201 Iv E JUN 271979

MAT. LAB.

NATIONAL COOPERATIVE HIGHWAY RESEARCH PROGRAM REPORT 201 ACCEPTANCE CRITERIA FOR ELECTROSLAG WELDMENTS IN BRIDGES

TRANSPORTATION RESEARCH BOARD NATIONAL RESEARCH COUNCIL

RFSEARCH TON I TRANSPORTATION RESEARCH BOARD 1979 Officers PETER G. KOLTNOW, Chairman THOMAS D. MORELAND, Vice Chairman W. N. CAREY, JR., Executive Director

Executive Committee HENRIK E. STAFSETH, Executive Director, American Assn. of State Highway and Transportation Officials (ex officio) LANGHORNE M. BOND, Federal Aviation Administrator, U.S. Department of Transportation (ex officio) KARL S. BOWERS, Federal Highway Administrator, U.S. Department of Transportation (ex officio) RICHARD S. PAGE, Urban Mass Transportation Administrator, U.S. Department of Transportation (ex officio) JOHN M. SULLIVAN, Federal Railroad Administrator, U.S. Department of Transportation (ex officio) HARVEY BROOKS, Chairman, Commission on Sociotechnical Systems, National Research Council (ex officio) ROBERT N. HUNTER, Chief Engineer, Missouri State Highway Department (ex officio, Past Chairman 1977) A. SCHEFFER LANG, Assistant to the President, Association of American Railroads (ex officio, Past Chairman 1978) HOWARD L. GAUTHIER, Professor of Geography, Ohio State University (ex officio, MTRB liaison) LAWRENCE D. DAHMS, Executive Director, Metropolitan Transportation Commission, San Francisco Bay Area ARTHUR C. FORD, Assistant Vice President (Long-Range Planning), Delta Air Lines ARTHUR J. HOLLAND, Mayor, City of Trenton, N.J. ROBERT R. KILEY, Management Analysis Center, Cambridge, Mass. JACK KINSTLINGER, Executive Director, Colorado Department of Highways PETER G. KOLTNOW, President, Highway Users Federation for Safety and Mobility THOMAS J. LAMPHIER, President, Transportation Division, Burlington Northern, Inc. ROGER L. MALLAR, Commissioner, Maine Department of Transportation MARVIN L. MANHEIM, Professor of Civil Engineering, Massachusetts Institute of Technology DARRELL V MANNING, Director, idaho Transportation Department ROBERT S. MICHAEL, Director of Aviation, City and County of Denver, Colorado THOMAS D. MORELAND, Commissioner and State Highway Engineer, Georgia Department of Transportation DANIEL MURPHY, County Executive, Oakland County, Michigan PHILIP J. RINGO, President, ATE Management & Services Co. MARK D. ROBESON, Chairman, Finance Committee, Yellow Freight Systems DOUGLAS N. SCHNEIDER, JR., Director, District of Columbia Department of Transportation WILLIAM R. SEARS, Professor of Aerospace and Mechanical Engineering, University of Arizona WILLIAM K. SMITH, Vice President (Transportation), General Mills JOHN R. TABB, Director, Mississippi State Highway Department JOHN P. WOODFORD, Director, Michigan Department of Transportation CHARLES V. WOOTAN, Director, Texas Transportation Institute, Texas A&M University

NATIONAL COOPERATIVE HIGHWAY RESEARCH PROGRAM

Transportation Research Board Executive Committee Subcommittee for the NCHRP PETER G. KOLTNOW, Highway Users Federation (Chairman) KARL S. BOWERS, U.S. Department of Transportation THOMAS D. MORELAND, Georgia Department of Transportation HARVEY BROOKS, National Research Council HENRIK E. STAFSETH, Amer. Assn. of State Hwy. and Transp. Officials A. SCHEFFER LANG, Association of American Railroads W. N. CAREY, JR., Transportation Research Board

Field of Materials and Construction Area of Specifications, Procedures, and Practices Project Panel D1O-10

G. J. HILL, Michigan Dept. of St. Hwys. & Transp. (Chairman) CARL E. HARTBOWER, Federal Highway Administration WARREN G. ALEXANDER, New York State Dept. of Transp PAUL JONES, California Dept. of Transportation ROBERT L. ANDERSON, Kansas Dept. of Transportation CLARENCE R. REA, Texas State Dept. of Hwys. & Public Transp. HARRY E. BROADBENT, JR., American Welding Society ROBERT W. SHARP, Lincoln Electric Company HARRY CZYZEWSKI, Metallurgical Engineers, inc. CHARLES H. MCGOGNEY, Federal Highway Administration KARL FRANK, University of Texas (Austin) W. G. GUNDERMAN, Transportation Research Board

Program Stafi KRIEGER W. HENDERSON, JR., Program Director LOUIS M. MACGREGOR, Administrative Engineer ROBERT E. SPICHER, Projects Engineer R. IAN KINGHAM, Projects Engineer HERBERT P. ORLAND, Editor ROBERT J. REILLY, Projects Engineer HELEN MACK, Associate Editor HARRY A. SMITH, Projects Engineer EDYTHE T. CRUMP, Assistant Editor NATIONAL COOPERATIVE HIGHWAY RESEARCH PROGRAM REPORT 201

ACCEPTANCE CRITERIA FOR ELECTROSLAG WELDMENTS IN BRIDGES

W. P. BENTER, JR., and C. G. SCHILLING UNITED STATES STEEL CORPORATION MONROEVILLE, PENNSYLVANIA

RESEARCH SPONSORED BY THE AMERICAN ASSOCIATION OF STATE HIGHWAY AND TRANSPORTATION OFFICIALS IN COOPERATION WITH THE FEDERAL' HIGHWAY ADMINISTRATION

AREAS OF INTEREST: STRUCTURES DESIGN AND PERFORMANCE CONSTRUCTION GENERAL MATERIALS (HIGHWAY TRANSPORTATION) (RAIL TRANSPORTATION)

TRANSPORTATION RESEARCH BOARD NATIONAL RESEARCH COUNCIL WASHINGTON, D.C. MAY 1979 NATIONAL COOPERATIVE HIGHWAY RESEARCH PROGRAM NCHRP Report 201

Systematic, well-designed research provides the most ef- Project 10-10 FY '74 fective approach to the solution of many problems facing ISSN 0077-5614 highway administrators and engineers. Often, highway ISBN 0-309-02905-8 problems are of local interest and can best be studied by L. C. Catalog Card No. 79-64880 highway departments individually or in cooperation with their state universities and others. However, the accelerat- Price: $5.20 ing growth of highway transportation develops increasingly complex problems of wide interest to highway authorities. These problems are best studied through a coordinated program of cooperative research. In recognition of these needs, the highway administrators Notice of the American Association of State Highway and Trans- portation Officials initiated in 1962 an objective national The project that is the subject of this report was a part of the National Cooperative Highway Research Program conducted by the highway research program employing modern scientific Transportation Research Board with the approval of the Governing techniques. This program is supported on a continuing Board of the National Research Council, acting in behalf of the basis by funds from participating member states of the National Academy of Sciences. Such approval reflects the Governing Board's judgment that the.program concerned is of national impor- Association and it receives the full cooperation and support tance and appropriate with respect to both the purposes and re- of the Federal Highway Administration, United States sources of the National Research Council. The members of the technical committee selected to monitor this Department of Transportation. project and to review this report were chosen for recognized The Transportation Research Board of the National Re- scholarly competence and with due consideration for the balance of disciplines appropriate to the project. The opinions and con- search Council was requested by the Association to admin- clusions expressed or implied are those of the research agency that ister the research program because of the Board's recog- performed the research, and, while they have been accepted as nized objectivity and understanding of modern research appropriate by the technical committee, they are not necessarily those of the Transportation Research Board, the National Research Coun- practices. The Board is uniquely suited for this purpose cil, the National Academy of Sciences, or the program sponsors. as: it maintains an extensive committee structure from Each report is reviewed and processed according to procedures which authorities on any highway transportation subject established and monitored by the Report Review Committee of the National Academy of Sciences. Distribution of the report is ap- may be drawn; it possesses avenues of communications and proved by the President of the Academy upon satisfactory comple- cooperation with federal, state, and local governmental tion of the review process. agencies, universities, and industry; its relationship to its The National Research Council is the principal operating agency of the National Academy of Sciences and the National Academy of parent organization, the National Academy of Sciences, a Engineering, serving government and other organizations. The private, nonprofit institution, is an insurance of objectivity; Transportation Research Board evolved from the 54-year-old High- way Research Board. The TRB incorporates all former HRB it maintains a full-time research correlation staff of special- activities but also performs additional functions under a broader ists in highway transportation matters to bring the findings scope involving all modes of transportation and the interactions of of research directly to those who are in a position to use transportation with society. them. The program is developed on the basis of research needs identified by chief administrators of the highway and trans- portation departments and by committees of AASHTO. Each year, specific areas of research needs to be included in the program are proposed to the Academy and the Board by the American Association of State Highway and Trans- portation Officials. Research projects to fulfill these needs are defined by the Board, and qualified research agencies are selected from those that have submitted proposals. Ad- ministration and surveillance of research contracts are Published reports of the responsibilities of the Academy and its Transportation Research Board. NATIONAL COOPERATIVE HIGHWAY RESEARCH PROGRAM The needs for highway research are many, and the National are available from: Cooperative Highway Research Program can make signifi- Transportation Research Board cant contributions to the solution of highway transportation National Academy of Sciences problems of mutual concern to many responsible groups. 2101 Constitution Avenue, N.W. The program, however, is intended to ccmplement rather Washington, D.C. 20418 than to substitute for or duplicate other highway. research programs. Printed in the United States of America.

FOREWORD This report contains the findings from an extensive laboratory investigation of electroslag weldments subjected to a variety of tests intended to assess performance By Staff that could be expected in actual bridges. The report is recommended to bridge Transportation engineers, researchers, and members of specification-writing bodies concerned with Research Board behavior of welded structures.

Electroslag welding offers some noteworthy economic advantages over other welding processes for certain types of bridge weldments. However, at the time this research was initiated (1974) some engineers were hesitant to permit use of the process for weldments subject to bridge loadings because of insufficient information on performance. The principal areas of concern were the physical and metallurgical properties of the weld and the heat-affected zones. Concerns about the adequacy of current electroslag welding specifications for bridges indicated a need for a thorough analytical and experimental program of evaluation of all parameters. The over-all objective of this project was to develop and verify acceptance criteria for the use of electroslag butt welds in bridges. Research was conducted in two phases. The specific objective of Phase I was to define necessary acceptance specifications based on an experimental study using laboratory specimens from full-size electroslag weldments of the type used in bridge girders. The specific objective of Phase II was to verify the findings of Phase I by conducting tests of full-size bridge girders under simulated service conditions. Phase I included preparation of a state-of-the-art report on electroslag weld- ing, which was published as NCHRP Research Results Digest 74, "Electroslag Weidments in Bridges." Variables in the Phase I specimens included welding procedures and equipment, grades of steel, and plate thicknesses. The testing was conducted to determine toughness, fatigue, tensile, and metallurgical properties. Tentative acceptance tests aid criteria for electroslag butt welds were developed based on the results of tests in Phase I. Flange butt welds were fabricated to obtain properties simulating these acceptance levels in full-size girders and were tested under simulated service conditions to determine fatigue. and fracture behavior in Phase II. This report contains the findings of both phases of research and includes suggested acceptance testing procedures for electroslag welding in bridges. More detailed information may be found in the individual final reports submitted by the research agency on each phase. Both are available on loan from the NCHRP Program Director. The specific titles are: "Acceptance Criteria for Electroslag Weidments in Bridges—Phase I" and 'Fatigue and Low-Temperature Impulse Tests of Electroslag Welded Girders—Phase II." Problems that developed in the field during the time this research was nearing completion have raised additional questions regarding the performance of electro- slag welds. As a result, the Federal Highway Administration in 1977 prohibited the use of electroslag welding in main structural tension members on Federal-aid projects and, at the same time, instituted a program of rigorous inspection (radio- graphic and ultrasonic) of nonredundant main load-carrying tension members in existing structures welded by the electroslag process. Because of these develop- ments, the findings of NCHRP Project 10-10 are not expected to lead directly to immediate increased use of electroslag welding in bridges; but much useful infor- mation on the process has been obtained and research needs have been clarified by this study. CONTENTS

1 SUMMARY 3 CHAPTER ONE Introduction and Research Approach Background State of the Art Objectives Experimental Work—Phase I Experimental Work—Phase II 17 CHAPTER TWO Findings Phase I Phase II 41 CHAPTER THREE Interpretation and Appraisal of Findings Background Electroslag Weidments Acceptance Criteria Recommendations and Future Work 43 References ACKNOWLEDGMENTS The research reported herein was performed under NCHRP ducted the investigation and prepared the section on thermal• Project 10-10 by the Monroeville, Pa., research laboratory of distribution. United States Steel Corporation with W. P. Benter, Jr., Asso- K. H. Klippstein made major contributions in the planning, ciate Research Consultant, and C. G. Schilling, Section Super- performance, and evaluation of the studies in Phase II. G. T. visor, as co-principal investigators. Blake, R. E. Droske, M. Humlan, C. E. Seigh, and R. M. Grateful acknowledgment is extended by these authors for Carratura assisted in planning and performing the girder tests. the important contributions of their colleagues, as follows: P. J. Konkol assisted in planning and evaluating welding pro- P. J. Konkol was the assistant principal investigator for cedures used in Phase II. R. F. McCartney and J. V. Pellegrino Phase I. The other authors of the Phase I report were B. M. performed failure-analysis studies on some of the details. Kapadia, who conducted the investigation and prepared the J. M. Barsom, A. K. Shoemaker, and J. M. Holt were responsi- section on fatigue crack-growth rate; A. K. Shoemaker, who ble for the fatigue-test program on the full-thickness electroslag- conducted the investigation and prepared the section on residual welded straight joints referred to in the text. L. G. Seigel made stresses; and J. F. Sovak, who conducted the investigation and valuable suggestions regarding the cooling system for the low- prepared the section on K1 testing. P. J. Konkol also con- temperature impulse tests of the girders. ACCEPTANCE. CRITERIA FOR ELECTROSLAG WELDMENTS IN BRIDGES I

SUMMARY The objective of NCHRP Project 10-10 was to develop acceptance criteria that can be used by highway engineers and transportation officials to assure that electro- slag weidments, when made by controlled procedures under production conditions, are suitable for use in tension members of bridges fabricated from approved struc- tural steels with minimum yield points up to 50 ksi (345 MPa). To meet this objective, in Phase I electroslag welds were produced in 1- and 4-in.-thick (25- and 102-mm) ASTM A36 and A588 Grade A steel plates (AASHTO M183 and M222), including transition joints. These were subjected to extensive evaluation by using a wide variety of recognized tests conducted at various test temperatures with specimens machined from standard (in accordance with AWS Dl. 1 Structural Welding Code) and nonstandard test locations. These tests included mechanical- property determinations, fatigue-crack growth-rate determinations, and a wide variety of fracture-toughness measurements. Supplemental studies of thermal distribution, residual stresses, chemical composition, macrostructure and micro- structure, soundness, and weathering characteristics were also conducted. These data were used in developing acceptance criteria. Instead of attempting to develop entirely new acceptance criteria, the AWS Dl. 1-75 Structural Welding Code was used as the basis for reference, and modifica- tions and restrictions to the code constitute the proposed tentative acceptance criteria. The data obtained showed that adequate strength and ductility across electro- slag weidments can be readily met with commercially available filler metals. The current AWS Dl. 1 testing procedures for all-weld-metal tensile properties and joint efficiency are adequate. The study showed that changing the type of welding shoe (solid copper vs water-cooled copper) can significantly affect tensile proper- ties. Because AWS Dl. 1 only requires requalification by radiography or ultrasonic testing when this essential variable is changed, the data showing the effect of a change in the welding shoe were presented to the Structural Welding Committee. The committee made the appropriate changes in AWS Dl. 1, which were published in the 1976 revision of the Code. Fatigue-crack growth-rate tests showed that in electroslag welds, the fatigue properties of the weld metal and associated heat-affected zones (HAZ) are as good as, or superior to, those of the base metal. In fact, considerable resistance to crack propagation was encountered during precracking (initial crack extension before obtaining crack-growth data) of the fatigue-crack-growth specimens. The results showed that adequate notch toughness for AASHTO Zone 1 and Zone 2 service can be obtained. The Charpy V-notch (CVN) test appears to be the most practical toughness test for production control purposes. CVN properties vary with location in the weld, both with respect to distance from the weld center- line and distance along the weld length. The location specified by AWS Dl. 1-75 for weld-metal test specimens (quarter-thickness position at weld centerline) appears to be the most representative and the best for control testing. Electroslag- weld-metal CVN test results showed more scatter than the test results for other weld metals; therefore, it is recommended that eight CVN specimens be tested (rather than 5 specimens as specified in AWS D1.1), that the highest and lowest values be discarded, and that the remaining six values be averaged. It is also 2 recommended that for Zone 1 and Zone 2 service, the CVN values average 15 ft-lb minimum at 0 F (20 J at —18 C), with no values 'less than 10 ft-lb (14 J), as required in Appendix C of AWS D1.1. For Zone 3 service under the present state of the art, electroslag weldments in the as-welded condition do not appear suitable for dynamic applications because of low HAZ toughness. In electroslag welds the HAZ is large and extends directly across the plate in butt joints; therefore, it is recommended that CVN tests (5 specimens at the quarter-point, discarding the highest and lowest values and averaging the three remaining values) be conducted on the grain-coarsened HAZ, with the notch locations determined by etching. For acceptance criteria, the AASHTO toughness requirements for the base plate should be met. For example, the HAZ of A36 steel for Zone 2 service must meet 15 ft-lb at 40 F (5 C). Minor indications of grain-boundary separations were noticed in a few welds and were investigated, even though they did not significantly affect strength, fatigue- crack-growth rate, and notch toughness. In qualification testing of electroslag welds, the presence of grain-boundary separations is best detected by the four side- bend tests required by AWS D1.1. This code does not permit any discontinuities over ½ in. (3.2 mm) in length after plastic straining (bending). The Phase II results cause some concern regarding the fatigue strength of electroslag weldments. These tests showed that electroslag-welded joints with a 1-in-21/2 taper in thickness cannot be safely included in AASHTO detail Category B when the weld reinforcement is removed and in Category C when the weld reinforcement is not removed. The present AASHTO fatigue specifications permit the use of these categories for straight and tapered joints welded by other processes. A review of available information suggests that electroslag-welded joints with a taper in thickness should have significantly shorter fatigue lives than straight electroslag-welded joints because a stress-concentration factor of about 2 is caused by the taper. A very limited amount of experimental information suggests that electroslag welds may have a somewhat higher fatigue notch sensitivity than submerged-arc welds. If this is true, the effect of stress concentration is somewhat greater for electroslag welds than for submerged-arc welds. There is little evidence to indicate that straight electroslag-welded joints with the weld reinforcement removed should not be classified as Category B detail or that such joints have a lower fatigue strength than similar joints made by other welding processes. More fatigue tests on full-thickness tapered and straight electroslag-welded joints in plate and girder specimens are needed to adequately define the fatigue strength of such joints. However, the present study suggests that Categories B and C would be appropriate for straight and tapered electroslag-welded joints, respec- tively, provided that the weld reinforcement is properly removed. The present study also suggests that electroslag-welded joints with the reinforcement in place should not be used in fatigue applications. The Phase II results showed that full-scale girder specimens can withstand- stresses at —30 F, which are well above the AASHTO maximum allowable bending stresses even when 20 percent or more of the tension-flange area is cracked. In service, a bridge member with a crack this size would soon fail by fatigue regardless of the level of fracture toughness in the joint. Thus, an increase in the fracture toughness of the joint would not significantly affect the service performance of such members. A large crack developed in one of the girder specimens at a nick caused by knocking off a tack-welded strongback. This crack did not affect the test results, but is mentioned here to emphasize the need for proper repair of any surface imperfections caused by removal of the strongbacks. 3

CHAPTER ONE

INTRODUCTION AND RESEARCH APPROACH

BACKGROUND To place this final report on NCHRP 10-10 in proper perspective and familiarize engineers who might not be Electroslag weldments have been used in many struc- currently involved in electroslag welding, salient features tures, including bridges, buildings, storage tanks, and of the state-of-the-art report (1) are included next; minor pressure vessels, and in components of structures that can updating has been done to cover developments since its be positioned so that the weld is made vertically. Where issuance. applicable, this process is often preferred because of eco- nomics and reproducibility of quality. Most applications STATE OF THE ART have been successful; however, in some instances (as with other welding processes) where the process has been mis- Electroslag welding is defined (Terms and Definitions, applied or the process variables have not been properly AWS A3.0-69, American Welding Society, 1969) as "a controlled, defective welds have been encountered. Some welding process wherein coalescence is produced by molten of these have caused serious problems and have raised slag that melts the filler metal and the surfaces of the, work doubt as to the consistency of weldments produced by to be welded. The weld pool is shielded by the slag that electroslag welding. As an interim measure, the Federal moves along the full cross section of the joint as welding Highway Administration (FHA) in early 1977 placed a progresses. The conductive slag is maintained molten by its ban on the use of electroslag weldments on main structural resistance to electric current passing between the electrode tension members on Federal-aid projects. This restriction and the work." Welding is done in the vertical or near- will continue until the quality of this type of weld can be vertical position, and joints are generally accomplished in a ensured by possible modification in the welding process single pass regardless of the thickness to be joined. and/or improvement in the inspection and quality-control Russian scientists from the Paton Institute of Electric procedures that appear necessary. In addition, the FHA Welding in Kiev are generally given credit for the develop- requested that nonredundant main load-carrying tension ment of the process, as it is known to'Jay, and the develop- members in existing structures that are known to have been ment of the necessary machines and techniques in the early welded with the electroslag process be subjected to rigorous 1950's (Welding Handbook, 6th edition, Chap. 48, nondestructive testing (radiographic and ultrasonic inspec- American Welding Society, 1970). By 1959, the process tion). The problems, when they do occur, have been was well developed and thoroughly described by the associated with cracks, extensive gas discontinuities, grain- Russians (Electroslag Welding, ed. B. E. Paton (rnslated boundary separations, and weld repairs. from Russian edition of 1959) American Welding Society, There has been a reluctance on the part of some engi- 1962). The first electroslag unit was introduced in the neers to permit the use of this relatively new process for United States in 1959. One of the first reported applica- weldments subject to loading of the type that occurs in tions of electroslag welding in bridge construction was in bridges, because sufficient research has not bccn conductcd 1965 (P. E. Masters and R. S. Zuchowski, "Vertical Sub- to determine their performance. merged-Arc Welding," Welding Journal, Oct. 1965, pp. To determine whether extensive use of this economic 829-837). Initial acceptance, as with many new welding tool in the fabrication of bridges is technically sound, a processes, was slow; however, its use increased rapidly. thorough analytical and experimental program was devel- The process is ideally suited for welding heavy plates oped. This program, designated. as NCHRP Project 10-10, that can be positioned in the vertical or near-vertical posi- "Acceptance Criteria for Electroslag Weldments in tion; this process becomes increasingly more economical as Bridges," was initiated May 1, 1974, and is divided into two plate thickness increases. The most common use appears to phases. Phase I was designed to define required acceptance be for plate thicknesses in the range 1 .to 8 in. (25 to specifications for electroslag welds by conducting a litera- 200 mm). Equipment is commercially available for butt- ture survey (1) and by evaluating the results of extensive welding plates as thin as ½ in. (12 mm) and up to 36 in. laboratory tests. Phase II was intended to verify the con- (91 cm). Special multiwire units can be designed for even clusions from Phase I 'by conducting simulated service thicker sections. tests of full-size bridge girders containing electroslag butt- Electroslag welding offers a number of advantages and welds. The Phase I and Phase II agency reports are avail- disadvantages compared with submerged-arc, shielded- able through the Transportation Research Board (2, 3). This report is the final report on NCHRP 10-10 and metal-arc, and flux-cored-arc welding for butt- and tee- includes the salient findings and conclusions of the study. joints of the type that are used in bridge and building The details and complete experimental data are included in construction. Some of these are listed in the sections that the previous reports (1, 2, 3). follow. 4

Advantages The weld HAZ is extremely large and much of it is coarse grained. The electroslag process, once started, is completely continuous (100 percent operating time). The weld is Variations in welding parameters can affect the degree completed in one pass; therefore, there is only one setup of admixture which, in turn, affects deposited weld metal and no need for interpass cleaning or repositioning of chemical composition. material or parts. The effects of the metallurgical characteristics, particu- Extremely high deposition rates of 35 to 45 lb (16 to larly the coarse grain size of the weld metal and the HAZ, 20 kg) per hour per electrode are achievable. have raised questions regarding the suitability of electro- Welding-material savings result because no stub or slag weldments for certain applications and the need for, or spatter losses occur, and flux composition is only about advantages of, heat treatment after welding to refine the 5 percent of the deposited-metal weight. grain structure. Joint preparation is minimized. Mill edges and square oxygen-cut edges are normally employed. Minor Principles of Operation gouges and cutting irregularities can be tolerated. Smooth, predetermined contours are produced on There are a number of variations of the electroslag the weld reinforcement. Postweld cleaning is minimal. process; however, all have common characteristics. The High quality, sound weld deposits are produced. process is initiated in a sump at the bottom of the joint. The weld metal stays molten longer, allowing gases to This sump is later removed and discarded, or at least is escape and nonmetallic inclusions to float to the slag above never considered for structural purposes a part of the the weld pool. - finished joint. The electroslag weld is initiated by striking Weld chemical composition can be controlled pre- an electric are beneath a layer of granular welding flux, cisely by adjusting welding consumables and parameters. like a conventional submerged-arc weld. Steel wool or Because of the symmetry of most vertical welds, iron powder is often used to aid in starting the arc. When. there is no angular distortion in the horizontal plane. There a sufficient volume of hot molten flux forms, the arc action is minimum distortion in the vertical plane, which can be stops and the current passes from the electrode to the easily controlled. workpiece through the electrically conductive slag. At this Preheating is generally not required. point, welding progresses quietly and spatter ceases. The Electroslag welding is the fastest welding process resistance heating of the molten slag is sufficient to melt the for large, thick joints. electrode and the edges of the workpiece. The temperature The process provides good operator comfort and of the slag is in the vicinity of 3500 F (1925 C). The ease of training. liquid metal from the electrode and from the molten edges For the foregoing reasons, electroslag welding is a very of the workpiece passes through the slag and forms a attractive process for making large welds in building and liquid-metal pool below the slag bath, which then solidifies. bridge constructions, as well as for other applications In this manner, the weld progresses vertically upward and requiringlarge heavy welds, and its use can result in sizable is usually terminated in a runoff tab above the structural reductions in fabricating costs. The advantages are partly joint. The end of the weld often contains crater cracks and counteracted, however, by process and metallurgical shrinkage voids that are removed, when the runoff tab is disadvantages. removed. Details of the current electroslag-welding practices and Process Disadvantages the necessary equipment, power supplies, and techniques are described in literature (4). All welds must be made in a vertical or near-vertical position. Materials Unless welding parameters are properly controlled, centerline cracking can occur because of unfavorable Base Materials dendrite orientation. Electroslag welding can and is being used for a wide If the welding process is discontinued for any reason, variety of carbon, high-strength low-alloy, alloy, and stain- restarting will generally result in. a major imperfection that requires repair. less steels and certain nonferrous metals. The AWS Structural Welding Code (5) recognizes electroslag welding as a suitable process for welding ASTM A36, A242, A441, Metallurgical Disadvantages A572 Grades 42 to 65, and A588 structural steels in all The very high heat. input of the process results in a thicknesses, with welds to be used in the as-welded condi- protracted thermal cycle with very slow solidification and tion. It specifically prohibits the use of the process on cooling rates. These have the following metallurgical quenched-and-tempered steels. . effects: There is general agreement that electroslag weldments in steels are suitable for all types of ferritic and martensitic 1. The grains in the weld metal are large and columnar. steels provided they are given a heat treatment to refine the They are oriented horizontally at the weld edges and turn structure in both the weld metal and the weld HAZ. Dis- to a vertical orientation at the center. agreement exists regarding the conditions under which electroslag weidments may be suitable for use in the as- 2. Butt joints: 1 in. to 1 in. (25 mm) welded or stress-relieved conditions. Many users are pro- 4 in. to4in. (102 mm) ceeding very cautiously in this area. 1 in. to 2 in. (51 mm) 2 in. to 4 in. Electrodes 3. Consumables: EH-14-EW, EM15K-EW, EWS- The wires used for electroslag welding are very similar EW, and EWT1 of AWS'A5.25-78 to those used for submerged-arc, gas-metal-arc, and flux- 4. Welding shoes: Water-cooled copper and solid cored-arc welding. In practice, many of the wires devel- copper oped for these older processes are being used for electro- slag welding. Solid, flux-cored, and metal-powder-cored 5. Welding process: Stationary and oscillating consuma- electrodes are being used. ble guide and oscillating wire 6. Weld length: 18- and 42-in. (458 and 1067 mm) Fluxes water-cooled shoes, 24-in. (610 The fluxes for electroslag welding are generally pro- mm) solid shoes prietary mixtures. The flux performs very critical functions even though the quantity used is relatively small (approxi- A total of 22 weldment types was made. Details of mately 5 percent by weight of the filler metal deposited). these weldments are given in Table 1. Resistivity, viscosity, and chemical activity must be con- trolled to ensure that the process functions properly and Materials that the deposited weld metal is sound and has the desired Steel Plates mechanical properties. For welding steels such as those used in bridges, the consumables are in accordance with The A36 and A588 Grade A plates used in the program AWS A5.25-78, "Specification for Consumables for Elec- were ordered directly from the producing mill and were troslag Welding of Carbon and High Strength Low Alloy produced by conventional commercial practices. The Steels." 4-in.-thick A36 plates had been normalized in accordance with mill practice to ensure that the AASH'I'O notch- Joint Geometry toughness requirement for Zone 2 service of 15 ft-lb (20 J) CVN energy absorption at +40 F (5 C) would be met. For butt joints, square-groove joint geometries are usu- All other plates were produced in the hot-rolled condition. ally employed. (Single V-joints are sometimes used on The chemical compositions of the steels, as determined thin plates welded by the conventional electroslag process.) by laboratory check analyses, are given in Table 2. The Transition joints are made by chamfering one or both sides tensile properties and the CVN impact properties are given of the thicker member to the appropriate slope. The weld in Tables 3 and 4, respectively. can be a butt joint as thick as the thinnest member, or the weld can be tapered to include part or all of the transition Electrodes and Consumable Guide Tubes slope. Special retaining shoes are available for transition joints. Four types of electrodes, all %2 in. (2.4 mm) in diam- Tee-welds are also a common application of electroslag eter, were used in the program. The chemical compositions welding. With the consumable-guide method, they are of the electrodes and guide tube are given in Table 5. The usually made in one pass with one or two electrodes de- guide tubes, 16-in. ID by 5/8-in. OD (4.76 by 15.88 mm), pending on thickness. Special shoes are available or can be are Si-Al-killed C-Mn steel with a thin flux coating on the fabricated with the desired fillet contour. outside.

OBJECTIVES Fluxes The over-all objective of this project is to develop and The two fluxes used in the program were those recom- verify acceptance criteria for the use of electroslag butt- mended, by the electrode manufacturers for use with their welds in bridges. The specific objective of Phase I was to particular electrodes. Flux L is a fused flux originally define the necessary acceptance specifications based on designed for submerged-arc welding but also recommended studies of laboratory specimens from full-size welds. The for electroslag welding (ESW) with.electrodes A, B, and objective of Phase II was to verify the findings of Phase I D. Flux H is a fused flux specifically designed for ESW by conducting simulated service tests on full-size bridge and recommended for use with electrode C. girders. Specimen Preparation EXPERIMENTAL WORK—PHASE I Joint Preparation Scope The A36 and A588 joints were prepared by oxygen- In Phase I, the following variables were studied: cutting, with a square bevel, 24-in.-wide (610-mm) pieces 1. Steel: ASTM A36 and ASTM A588 by the desired plate length. Sumps and runoff tabs were Grade A also cut from the same plate material. 6

TABLE 1 DESCRIPTION OF WELDMENTS INVESTIGATED Plate Weld Copper Weidment Steel Thickness, Length, Electrode Shoe No. Grade inches inches Type* Wire Feed Type

1 A36 1-1 42 A Stationary tube Water-cooled 2 A36 1-1 42 A Oscillating wire Water-cooled 3 A36 4-4 42 A Oscillating tube Water-cooled 4 A36 4-4 42 A Two stationary tubes Wáer-cooled 5 A36 4-4 42 A Oscillating wire Water-cooled 6 A588 1-1 42 C Stationary tube Water-cooled 6A A588 1-1 18 C Stationary tube Water-cooled 7 A588 1-1 42 C Oscillating wire Water-cooled 8 A588 4-4 42 C Oscillating tube Water-cooled 8A A588 4-4 18 C Oscillating tube Water-cooled 9 A588 4-4 42 C Two stationary tubes Water-cooled 10 A588 4-4 42 C Oscillating wire Water-cooled 11 A588 1-2 42 C Stationary tube Water-cooled 12 A588 2-4 42 C Oscillating tube Water-cooled 13 A36 1-1 24 A Stationary tube Solid 14 A588 1-1 24 C Stationary tube Solid 15 A588 4-4 24 C Oscillating tube Solid 16 A588 4-4 24 C Two stationary tubes Solid 17 A36 1-1 42 B Stationary tube Water-cooled 18 A36 4-4 42 B Oscillating tube Water-cooled 19 A588 1-1 42 D Stationary tube Water-cooled 20 A588 4-4 42 D Oscillating tube Water-cooled * A : EWT1, cored. Conversion Factor: B : EH14-EW, solid C : EMI3K-EW, solid - 1 inch = 25.4 mm D : EWS-EW, solid

TABLE 2 CHEMICAL COMPOSITION OF PLATES Plate Thickness, Steel Grade inches C Mn P S Si Cu Ni Cr Mo V Ti Al N 0

A36 1 0.21 1.16 0.011 0.024 0.03 0.04 0.03 0.06 0.01 <0.005 - <0.002 0.004 0.0101

A36 4 0.25 1.08 0.010 0.025 0.24 0.02 0.03 0.03 0.01 <0.005 - 0.032 0.005 0.0025

A588 grade A 1 0.15 1.12 0.013 0.029 0.24 0.34 0.05 0.65 0.02 0.05 0.005 0.036 0.005 0.0017

A588 grade A 2 0.14 1.09 0.015 0.028 0.21 0.29 0.15 0.64 0.01 0.04 0.005 0.042 0.004 0.0018

A588 grade A 4 0.19 1.18 0.009 0.019 0.25 0.32 0.26 0.55 0.06 0.06 - 0.032 0.005 0.0017 Conversion Factor:

1 inch = 25.4 mm

The transition joints (1 to 2 in. and 2 to 4 in.) were Joint Fixturing prepared for welding by oxygen cutting a chamfer with a The joint gap was 11/8 in. at the bottom and 11/4 in. slope of 21/2 to 1 on one side of the thicker plate (offset (28.6 and 31.8 mm) at the top for all weldments regardless alignment), as shown in Figure 9.20.1 of the AWS Struc- of weld length. The plates were held in position by either tural Welding Code (5). For the present program, both two or three strongbacks depending on weld length. the 4-in.-thick plate and the weld surface were part of the chamfer. All plates were ultrasonically inspected adjacent Welding to the joint face prior to welding to ensure that plates with indications that would be rejectable when the weld was No preheat was used and all plates were at room tested in accordance with the Structural Welding Code temperature at the time welding was started. When water- would not be used. cooled copper shoes were used, the water was not turned 7

TABLE 3 TENSILE PROPERTIES OF PLATES *

Plate Elongation Tlaiukgiess, Test Yield strength Tensile in 2 Inches, Reduction Steel Grade inches DirectiOn (0.2% Offset), Isi Strength, ksi 1 of Area, %

A36 1 L 41.4 75.0 30.5 67.2

T 39.3 70.4 30.0 61.6

A36 4 L 39.5 74.7 30.0 62.8

T 40.7 74.4 29.0 55.7

A588 1 L 62.6 94.2 23.5 63.0

T 55.6 84.2 21.2 49.4

A588 2 L 515 79.1 28.8 71.8

T 52.0 80.7 24.0 50.6

A588 4 L 54.4 87.9 24.5 67.1

T 55.3 88.5 22.5 56.2 * Average of duplicate 0.505-inch-diameter specimens Conversion. Factors: L = longitudinal or rolling direction 1. inch = 25.4 nun 1 ksi = 6.894 MPa T = transverse to rolling direction

on until welding commenced; the outlet water temperature thermocouples at various locations in the base metal so that was monitored and the flow, rate adjusted so that the outlet thermal histories could be recorded. temperature did not exceed 150 F (65 C). For each type of weldment, welding parameters that Macroscopic Examination were. judged to result in sound welds with suitable bead appearance were selected by the fabricator. Trial joints Full cross sections through the weld and HAZ were were often made to optimize these parameters. For most cut from each type of weldment in three orientations: (1) types of weldments, more than one joint was needed to normal to the weld axis, (2) parallel to the weld at the obtain sufficient material for the various mechanical- centerline of the joint, and (3) parallel to the plate surface property tests. With few exceptions, the parameters were at the midthickness of the plate. On selected weldments, held constant for each set of joints so that material from transverse sections were cut from the extreme top and any joint would be representative of all joints for a given bottom of the joints and were compared with similar sec- type of weldment. tions of the weld away from the ends. Planimeter measure- The conditions used to fabricate each weldment are given ments of the various regions of the weld metal and HAZ in Table 6. Variations in fill rate, flux consumption, and were made from the transverse sections, including those at energy input could have been caused by normal variations the extreme top and bottom of welds. in joint fit-up, shoe contact, and shrinkage strains during welding. Residual-Strain Measurements Representative A588 steel weldments were selected to Nondestructive Inspection conduct residual-strain measurements. On all weldments All welds were radiographed and ultrasonically tested investigated, the following measurements were made: (1) in accordance with the portions of AWS Dl.!-75 applicable the shrinkage across the weld after welding, (2) the to bridges. Only weldments or portions of weldments that residual transverse strain (normal to the weld axis) in the met the requirements of the aforementioned code were HAZ and the base plate, and (3) the longitudinal strain used for further testing. (parallel to the weld axis) adjacent to the deposited weld. The completed weldments were also photographed to Attempts were also made to determine the through-thick- record the heat-tint pattern and any irregularities in bead ness residual stresses in a few weldments. appearance. Microscopic Examination Thermal Distribution On selected and representative weldments, detailed To study the effects of several process parameters on the metallographic examinations were conducted. Grain- thermal distribution and subsequent mechanical properties boundary fissures or separations when noted were examined of EW weldments, eight weldments contained embedded in detail. 8

Hardness Traverses exposed to a semi-industrial atmosphere and subjected to weathering to evaluate the effect of electroslag welding on The previously mentioned sections normal to the weld, the general corrosion and appearance of the weldment. including those at the ends of welds, were used to determine the hardness distribution across the weld including subsur- Mechanical-Property Tests face, quarterpoint, and midpoint Rockwell B hardness traverses. Tension Tests Duplicate longitudinally oriented all-weld-metal standard Chemical Composition 0.505-in.-diameter (12.8-mm) tension-test specimens were The chemical composition of the weld metal, including taken from the midwidth of the weld of each type of trace elements, was determined for each type of weldment. weldment. For 1-in.-thick weidments and transition weld- Sampling location was at the midthickness of the weld ments, the test specimens were located at the midthickness; metal. for 4-in.-thick weldrnents, the specimens were located near the surface, at the quarter thickness, and at the midthick- ness of the weld metal. Weathering Tests In addition, duplicate full-thickness transverse strap- Sections of each of the ASTM A588 weldments were tension tests were conducted on each type of weldment.

TABLE 4- INDIVIDUAL CHARPY V-NOTCH IMPACT-TEST RESULTS ON PLATES Plate Test Energy Absorbed, Lateral Expansion, Shear Steel Thickness, Temperature, ft-lb mils Fracture, % Grade inches °F L T L T L T

A36 1 200 116,116 62,64 78,74 61,60 100,100 100,100 160 118,115 61,62 80,80 60,61 99,100 99,99 120 120,117 58,54 80,82 59,53 95,90 95,90 75 109,95 45,38 78,74 46,42 80,75 65,60 60- 43 37 35 - 32 50 45 40 20,49,82 22,17,26 25,46,68 27,22,31 20,35,60 35,30,35 0 11,23 8,11 14,24 12,14 10,15 10,10 -40 4,4 3,3 3,4 3,3 5,5 5,5

A36 4 125 - 60,56 - 62,60 - 100,100 100 80,83 54,56 72,72 59,58 100,100 99,99 75 73,70 56,57 69,63 55,56 90,85 90,90 55 58,56 45,46 55,54 45,45 65,70 - 70,65 40 66,41,49 31,30,34 60,39,46 32,34,36 60,45,50 50,50,50 20 37,42 32,29 36,39 33,30 50,45 40,35 0 39,33,14 22,21,22 36,31,17 23,21,24 40,30,20 25,25,25 -20 21,29 - 20,26 - 15,20 -

A588 1 140 80,76 65,68 100,100 120 74,80 63,64 100,100 100 53,53 47,48 70,65 75 39,40 37,38 40,40 60 38,40 35,35 30,35 40 23,33 21,27 10,15 0 8,16 7,13 5,5 -40 6,7 4,5 5,5 A588 2 160 105,91 33,35 76,73 40,42 100,100 99,99 120 105,89 32,35 75,72 37,38 100,98 95,90 75 105,100 25,19 74,73 28,24 100,85 50,45 60 52,72 18,21 45,60 22,25 35,55 20,35 40 40,51,53 18,22,19 33,43,44 20,24,21 15,20,20 25,30,30 0 12,9 11,8 13,10 9,7 10,5 5,5 -40 5,11 11,5 3,9 9,3 5,5 5,5

A588 4 200 91,91 61,64 100,100 160 65,93 50,63 70,100 120 52,56 41,42 50,60 100 48 41 40 75 30,35 26,31 15,20 40 31,38,38 24,29,30 10,10,10 0 • 12,15 11,11 5,5 -40 10,9 7,6 5,5 Conversion Factors:

1 inch = 25.4 mm = 5/9(°F - 32) 1 ft-lb = 1.356 J 1 mil = 0.0254 mm TABLE 5 CHEMICAL COMPOSITION OF ELECTRODES AND GUIDE TUBE-PERCENT Elec- AWS trode C14ification C Mn P S Si ( ii Ni Cr Mn 7 Ti Al N As lb On 0

A** EWT1 0.03 1.71 0.011 0.021 0.27 0.02 0.04 0.02 0.01 0.005 0.005 <0.002 0.005 - - - 3.2

E1114-EW 0.12 1.69 0.007 0.022 0.03 0.16 0.04 0.05 0.01 0.005 - 0.004 0.004 - - - 6.3

EM13K-EW 0.11 1.08 0.021 0.019 0.48 0.03 0.04 0.02 0.01 0.005 0.005 <0.002 0.006 0.044 0.0005 0.002 14.1

EWS-EW 0.11 0.55 0.006 0.016 0.30 0.57 0.56 0.63 0.02 <0.005 <0.005 0.012 0.008 - - - 5.6

Guide Tube 0.14 0.75 0.003 0.021 0.21 0.02 0.02 0.02 <0.01 .0,005 0." )5 0.0660.005 - * Parts per million. ** Analysis of Elentrode A was obtained or. an undilu"

Fatigue-Crack Growth-Rate Tests weldments. Specimen preparation and testing were in accordance with ASTM E208-69. For the 1-in.-thick weld- Fatigue-crack growth-rate studies were conducted on six ments, type P-i specimens (1 by 31/2 by 14 in., or 25.4 by 1-in.-thick and four 4-in.-thick representative butt-welds. 89 by 356 mm) were used and oriented with the broad face on the plate surface. Three sets were prepared from each Charpy V-Notch Tests 1-in.-thick weldment. One set had the hard-facing-bead The CVN energy absorption of the weld metal and notch at the center of the electroslag weld, the second set HAZ was determined. For the 1-in.-thick joints, transverse had the hard-facing-bead notch in the grain-coarsened CVN specimens were taken from, the midthickness and region of the HAZ, and the third set was obtained from the midwidth of the weld joint (notch perpendicular to the base plate. Drop-weight nil-ductility transition (NDT) plate surface), and transverse CVN specimens were taken temperatures for each zone were determined. from the midthickness of the plate in the grain-coarsened For the 4-in.-thick weldments, type P-3 specimens region of the HAZ with the notch perpendicular to the plate ( 5/8 by 2 by 5 in., or 15.9 by 51 by 127 mm) were used surface. For the 4-in.-thick joints, transverse CVN speci- and oriented in two directions. One set had the broad mens were taken from the surface, quarter thickness, and face at the weldment surface with the hard-facing bead midthickness at the midwidth and at the quarter width of placed on the weld or plate surface; the other set had the the weld metal, and transverse specimens were taken from broad face normal to the plate surface so that the crack the quarter of the plate in the grain-coarsened region of propagated along the weld axis (one side of the specimen the HAZ as previously described. In the transition joints, was the plate surface). Six sets were prepared for each the weld-metal CVN specimens were obtained from the weldment, so that weld metal, grain-coarsened HAZ, and approximate midwidth and midthickness location and the base plate were tested in two orientations. HAZ CVN specimens were obtained at the quarter- thickness location for the 2- and 4-in.-thick plates and at Dynamic-Tear Tests the midthickness location for the 1-in.-thick plate, with the notch in the grain-coarsened region. Dynamic-tear tests were conducted on the 4-in.-thick For each location (six locations in the weld metal and A36 and A588 base metals, on the weld metal in five one in the HAZ) in each type of weldment, 14 CVN weldments, and on the HAZ in three weldments. The specimens were prepared when sufficient material was specimen dimeisions were 7.125 by 1.60 by 0.625 in. (181 available. Because 0 F (-18 C) was the test temperature by 40.6 by 15.9 mm), with a pressed tip at the root of of most interest, five specimens were tested at 0 F and the the notch. The specimens were oriented transverse to the results were reported by using the AWS-recommended weld at the quarter thickness location, and the direction of method of eliminating the highest and lowest values and crack propagation was parallel to the weld axis. averaging the three remaining values. Three specimens each were tested at -20, 40, and 75 F (-30, 5, and 25 C). Five CVN specimens were obtained from the prescribed K1< Tests locations in the weld metal at the extreme bottom (start) Fracture-toughness (K1< ) tests were conducted on four and top (finish) ends of selected weldments and tested weldments. Two of the weldments were 1-in, to i-in, butt at 0 F. Similarly, five specimens were obtained from the joints, and the remaining two were 4-in, to 4-in, butt joints. HAZ at the bottom and top ends of weldments and tested The test specimens were 1 in. thick for the 1-in, butt joints at 0 F. and 2 in. thick for the 4-in.-thick butt joints. The fatigue cracks were oriented perpendicular to the plate surface and Drop-Weight Tests parallel to the weld in both the 1- and the 4-in.-thick weld- Drop-Weight-test specimens were prepared from four ments, and perpendicular to the plate surface and parallel 10

TABLE 6 WElDING CONDITIONS Wire- Curreut Plate Weld Filler- Flux Feed per Fill Energy Weldment Steel Thickness, Length, Shoe Electrode Metal Flux Consumed, Speed, Electrod, Voltage, Rate, Input, No. Grade inches inches Feed** Type+ Type grams ipm amperes volts ipe kJ/inch

1-1 A36 1 42 W. 1C A L 200 122 500 38 0.750 1520

1-2 A36 1 42 WC 1C A L 390 122 500 38 0.764 1740

1-3 A36 1 42 WC IC A L 300 122 500 37 0.768 1445

1-4 636 1 42 WC IC A L - 122 500 38 - -

2-. 6 1 42 (4(2 €34 A L 680 - 450 39 1.33 790

22 A36 1 42 WC OW . A L 340 - 500 .40 1.07 1122 3-1 636 4 42 (8 OC A L 400 244 650 50 0.409 4760

1-2 636 4 42 (4(2 A L 300 244 650 50 0.400 4860

3-4 636 4 42 (4' . 1X2 A L 350 244 650 50 0.379 5150

3-4 638 4 42 WC 02 A L 400 244 650 50 0.414 4710

3-6 .636 4 42 WC OC A L 300 244 650 50 0.383 5080

3-7 A3, 4 20 We 1X 1, L 300 244 650 50 0.333 5860

4-1 A36 4 42 WC 21 A L 360 135 450 44 0.484 4920

5-1 6,6 4 42 We OW A 1. 680 - 500 50 0.396 3785 6-1 A588 42 WI 1C C H 110 139 500 37 0.885 1255

6-2 A588 1 42 WC, IC C H 280 140 500 37 0.799 1390

6-3 6588 1 42 W" 10 C H 110 143 500 37 0.919 1210

6-4 6588 1 42 WC: 10 C H 116 143 500 37 0.900 1235 66-1 6188 1 11 JO C H 180 141 500 37 0.846 1310

o,.-2 20 1 lb WC 1(2 C H 75 139 500 37 0.900 1235 7-1 A588 1 4. e OW C H - - 500 35 1.28 821

1 42 OW C H 340 - 500 40 1.48 810 8-1 .5503 4 42 WC X C H 530 244 €50 51 0.429 4640

42 WC .: C H 500 240 650 50 0.409 4760

8-3 6588 4 42 WC .5: C H 800 244 650 50 0.435 4480

8-4 A588 4 42 WC 00 (2 H 500 244 650 51 0.409 4860

8-5 6588 4 19 WC 00 C H 575 244 650 50 0.358 5450

86-1 A588 4 18 WC 00 C H .400 244 650 50 0.480 4060

86-2 6588 4 18 WC OC C H 400 244 650 50 0.396 4930

9-1 A588 4 42 WC 2C C H 600 135 450 44 0.505 4700

9-3 6588 4 42 WC 2C C H 600 135 450 44 0.543. 4370

to the weld and running in the through-thickness direction EXPERIMENTAL WORK—PHASE II only in the 4-in.-thick weidments. Test Plan Duplicate specimens were tested for each of the afore- mentioned conditions, resulting in a total of 36 specimens. Tests were performed on two girder specimens, one of All the slow-bend K3 tests were conducted in duplicate A36 steel and the other of A588 steel. These specimens at —30 F (-35 C) and at an intermediate rate of loading are subsequently referred to as specimens 36 and 588, that corresponded to about 1 sec to failure. respectively. Each specimen contained two electroslag- 11

TABLE 6 (Continued) Wire- Current Plate Weld Filler- Flux Feed per Fill Energy Weldrnent Steel Thickness, Length, Shoe Electrode Metal Flux 'Consumed, Speed, Electrode, Voltage, Rate, Input, No. Grade inches inches Type* Feed** Type+ Type grams ipa amperes volts ipa kJ/inch

10-2 P.588 4 42 WC OW C H - - 500 51 0.417 3665

11-1 P.588 1-2 42 WC lC C H 280 147 500 47 0.715 1970

11-2 P.588 1-2 42 WC 1C C H 300 147 500 47 0.676 2080

12-1 P.588 2-4 42 WC 0C C H 440 144 500 47 0.420 3360

12-2 A588 2-4 42 WC OC C H 300 144 500 47 0.409 3445

13_1 P.36 1 24 S 1C A L 125 122 500 38 0.789 1450

13-2 P.36 1 24 5 1C A L 185 122 500 37 0.789 1410

13-3 A36 1 24 S 1C A L 375 122 500 38 0.765 1490 14-1 P.588 1 24 S 1C C H 80 143 500 37 0.824 1350

14-2 P.588 1 24 S 1C C H 180 140 500 37 0.916 1210

14-3 P.588 1 24 S 1C C H 190 139 500 37 0.946 1170

15-1 P.588 4 24 S OC C H 380 244 650 52 0.435 4660

15-2 P.588 4 24 5 CC C H 600 244 650. 50 0.388 5025

15-4 P.588 4 24 S CC C H 300 244 650 SO 0.372 5240

16-1 P.588 4 24 S 2C C H 280 189 600 44 0.697 4540

16-2 P.588 4 24 S 2C C H 450 135 450 44 0.520 4560

16-4 P.588 4 24 S 2C C H 350 135 450 44 0.540 4400

16-5 P.588 4 24 S 2C C H 350 , 135 450 44 0.578 4110

16-6 A88 4 24 5 2C C H 350 135 450 44 0.540 4400 17-1 A36 1 42 MC 1C B H 133 145 500 37 0.920 1205

17-2 P.36 1 42 WC 1C B L 115 146 500 37 0.959 1160 18-1 P.36 4 42 WC CC B L 500 244 650 50 0.391 5000

18-2 A36 4 42 WC CC B L 400 244 650 50 - -

18-3 A36 4 42 WC CC B L 400 244 650 50 0.408 4780

19-1 P.588 1 42 WC 1C D 1 105 143 525 37 0.979 1190 I 19-2 P.588 1 42 MC 1C D L 180 143 525 37 1.02 1140

20-1 P.588 4 42 MC OC D L 700 244 650 51 0.415 4800.

20-2 P.588 4 42 WC CC 0 1 700 244 650 51 0.406 4900

* WC : Water-cooled copper shoes. Conversion Factors: S : Solid copper shoes. ** lC : Single fixed consumable guide tube. i inch 25.4 nan 2C : Two fixed consumable guide tubes. 1 iptn = 0.423 sun/s CC : Single oscillating consumable guide tube. 1 kJ/inch = 0.0394 kJ/mzn OW : Single oscillating wire (nonconsumable guide tube). + See Table 5. welded transverse joints between 2-in.-thick and 4-in.-thick B details according to the latest AASHTO fatigue speci- tension-flange plates. It was decided to test tapered joints fications (6). One of the joints in specimen 588 had the because they are more severe than joints without a taper, weld reinforcement ground off, but the other did not and, although both types are included in the same detail category therefore, qualified as a Category C detail. For conveni- for fatigue design. ence in subsequent discussions, the joints in specimen 36 Both joints in specimen 36 had the weld reinforcement will be designated detail 36BN and detail 36ES. The ground off to a smooth contour and qualified as Category letter "B" is used to indicate that the joints, are Category 12

B details. Letters "N" and "S" are used to indicate the Specifically, the AASHTO allowable stress ranges of 18 north and south joints, respectively. The north and south and 13 ksi were applied to the Category B and Category C joints in specimen 588 are designated detail 588B and details, respectively. It was intended to apply these cyclic detail 588C, respectively. The letter "C" is used to indicate stresses for 2 million cycles—the design life corresponding that the joint is a Category C detail. The letter "R" will to these allowable stress ranges. be used to designate a repaired joint. Following the cyclic loading to 2 million cycles, the As shown in Figures 1 and 2, each specimen was first specimens were to be cooled to about —30 F (-34 C), the cyclically loaded at room temperature to simulate the minimum service temperature for AASHTO Zone 2 (6), traffic loading that occurs during the life of the bridge. and two impulse loads were to be applied. (An impulse is

DETAIL B B

If t = 304.8mm 2cr2' I ksi = 6.895 MPa Ii II °Cr5/9(°F-32) L 2 8' 2' U) U) 0 ksi FATIG U / 20kSi LI TEST ir- U) Ks, TIME STRESS DIAGRAM 1.5 Sec DETAIL B . B STRESS/TIME PLOT 242' -30°F -31°F -60°F

II I II 36ks1 36 ksi

12 2 ,'J 8' _,, U) U) LI I 20ksi

TEST 6.67ks' ksi L 1 4.85cc 4.8Sec Sec J TIME

STRESS DIAGRAM STRESS/TIME PLOT (IMPULSES I AND 2) Figure 1. Loading on specimen 36.

DETAIL B

I ft = 304.8mm I ksi = 6.895 MPa I I °C 5/9(°F-32)

U) FATIGUE 27 ksi 27 ksi

Seec ec TIME STRESS DIAGRAM STRESS/TIME PLOT (DETAIL B) DETAIL

12 8' L 1 27 ksi 27 ksi IMPULSE TEST ksi 9ksi

TIME STRESS DIAGRAM STRESS/TIME PLOT (IMPULSES lAND 2) Figure 2. Loading on specimen 588. 13 a load applied for a specified period of time and then TABLE 7 removed; in these tests, the load is varied from a minimum CHRONOLOGY OF TESTS to a maximum and then back to the original minimum.) Specimen 588 (Details 588B and 588C) The first was intended to simulate a normal service load. (1) Cyclic loading at room temperature; stress on Detail 588B Both joints were to be loaded to the AASHTO maximum varied from 9 to 27 ksi; stress on Detail 588C varied from allowable stress for steel used in that specimen-20 ksi 10 to 23 ksi. (138 MPa) for A36 steel and 27 ksi (186 MPa) for A588 (a) At 576,000 cycles, first crack observed in Detail 588C (b( At 617,000 cycles, crack in Detail 588C covered about steel. The second impulse load was intended to simulate 25 percent of flange area; crack repaired

a severe overload causing stresses equal to the specified (C) At 832.000 cycles, first crack observed in Detail 589B minimum yield strength of the material. To simulate dead (d) At 1,243,000 cycles, crack in Detail 5888 covered about 50 percent of flange area; no crack observed load, the impulse loads were superimposed on a constant in Detail 588CR minimum load corresponding to one-third of the maximum allowable static stress. (2) Sinusoidal impulse loadings applied at low temperatures. At 0'?, stress varied from 9 to 27 ksi; no failure or The stress rates for the impulse tests were chosen to visible damage approximate those observed in field measurements of At -30'!', programssed stress varied from 9 to 27 ksi; bridges under traffic (7). In these measurements, the total fracture at Detail 5888 at stress of 22 ksi impulse-load time for stress ranges of about 6 ksi (41 Specimen 36 (Details 3689 and 3685) MPa) was approximately 1 sec. Thus, the average strain (1) Cyclic loading at room temperature; stress on Details 368N and 36BS varied from 2 to 20 ksi. rate (that is, the stress range divided by the time to reach (a) At 448,000 cycles, first crack observed in Detail 36BS the peak stress) is 12 ksi/sec (83 MPa/sec). At 621.000 cycles, crack in Detail 368S covered about The actual testing program was somewhat different 55 percent of flange area; crack repaired from that planned, because the specimens developed sub- At 2,000.000 cycles, no crack observed in Detail 3688; 3/8-inch edge crack in Detail 36858 detected by stantial cracks before loading to 2 million cycles. Never- magnetic-particle inspection

theless, the basic plan of fatigue loading followed by low- (2) Sinusoidal impulse loadings at low temperatures temperature impulse tests was retained, as indicated by the (a) At -30'F, stress varied from 6.7 to 20 ksi; no failure chronology of the actual tests given in Table 7. or visible damage (b( At -30'F, stress varied from 6.7 to 36 ksi; no failure or visible damage Specimens (3) Cyclic loading at room temperature; stress on Details 3688 Figure 3 shows the nominal dimensions of the two and 368SR varied from 2 to 20 ksi. (a( At 2,442.000 cycles, crack in Detail 368SR covered specimens; details of the joints are also shown. The about 5.0 percent of flange area; no cracks observed nominal slope of the taper is 1 to 21/2, the maximum per- in Detail 3688; artificial crack covering about mitted for Categories B and C (6). The stress-concentra- 20 percent of flange area placed in Detail 368N (b) At 2,447,000 cycles, crack in Detail 368SR covered tion factor for detail 36BN was determined by strain gages iiut 3.4 percent of flange area; crack in Detail 3688 to be about 2. The mechanical properties and chemical covered about 20 percent of flange area compositions of the plates used in the specimens are given (4) Sinusoidal impulse and fatigue loadings at low temperatures (a) At -30'F, impulse stress varied from 6.7 to 20 ksi; no in Tables 8 and 9, respectively. visible damage The American Bridge Division of U.S. Steel Corporation (b) At -30'!'. 2000 cycles of fatigue stress varying from fabricated the specimens by using normal bridge-fabrication 2 to 20 ksi procedures (5, 8). The transverse tensionfiange joints (c) At -30'!', impulse stress varied from 6.7 to 36 ksi; no visible damage were made in the following way. First, the 4-in.-thick At-60'!', impulse stress varied from 6.7 to 36 ksi; no plate was tapered by gas cutting. Then, the electroslag visible damage welds were made by using the consumable-guide-tube At -60'!', 2500 cycles of fatigue stress varying from 2 version of the process, water-cooled copper shoes, and to 20 ksi; no visible damage )f( At -60'!', impulse stress varied from 6.7 to 36 ksi; one stationary guide tube. AWS EM13K-EW electrodes no visible damage

were used in specimen 36 and AWS EWS-EW electrodes (5) Cyclic loading at room temperature: stress on Details 368N were used in specimen 588. The welds were radiographi- and 36BSR varied from 2 to 20 ksi. cally inspected. These materials and procedures were (a) At 2,528,000 cycles, crack in Detail 36BSR covered about 30 percent of the flange area; crack in selected to provide CVN toughness values for the weld Detail 3658 covered about 37 percent of flange area

metal that are close to the minimum that would be per- )6( Sinusoidal impulse and fatigue loadings at low temperatures mitted by the acceptance criteria proposed in Phase I for (a) At -30'F, impulse stress varied from 6.7 to 20 ksi; Zone 2 (2); namely, 15 ft-lb at OF. Table 10 lists the no visible damage (b( At -30'?, 2000 cycles of fatigue stress varying from notch-toughness results for weld-metal specimens taken 2 to 20 ksi from extra material outside the finished dimensions of the (c( At -30'?, programmed impulse stress varied from 6.7 flange. The welding procedures were qualified in accord- to 36 ksi; fractured at Detail 36NSR at stress of

ance with the AWS Structural Welding Code (5), and the Notes, Listed Stresses and temperatures are nominal values. results of the required tests are given in Tables 11, 12, and Listed cyclic counts are cumulative counts from the start of the test of each specimen. No visible lamage 13. refers to the effect of the particular cycle of loading under consideration. 1 kSi = 6.895 MPa 'C = 5/9('F - 32) Test Setup and Procedures Sketches of the loading arrangements for various tests are 14

Detail

Specimen 36 6 B (Tension Flange Specimen 588 C B 12x 2' At Ends South Weld Detail-.... North 12"x4" At Center '1 (Web T-'z I L28'xi/2" (Compression 6-Il Specimen 36 13-10-1/2 -l/4 • l44-I/8" 11 Flone 114' x 2" i-6-I/16" 6-I/I Specimen 588 4-0-I/8 - 4-0-5/8 13-11-3/16"

WELDED BEAM

I ft *0.305 m in. *25.4mm

..._nforcement Off--- Weld Detail C Weld Reinforcement On -

WELD DETAIL

Figure 3. Test specimens.

TABLE 8 MECHANICAL PROPERTIES OF PLATES

Tests Yield Tensile Elongation CVR Energy Absorption Thickness, Heat Performed Strength, Strength, in 2 Incb. at4O*F, ft-lbs Component in. Steel No. At ksi ksi % 1 2 3 Mean

Flange 2 A588 748120 Mill 60.4 87.2 24.0 42 38 35 38 Flange 2 A588 748120 Lab 52.2 81.8 28.8 48 88 70 69 Flange 4 A588 72E579 Mill 58.2 90.8 23.0 27* 29* 25*, 27* Flange 4 A588 72E579 Lab 51.4** 85.1 20.5 48 39 35 41 Web 1/2 A588 658225 Mill 57.1 79.7 22.0 79 67 112 86 Specified Mm -A588 - - 50 70 21 13.3 13.3 13.3 20 Flange 2 A36 68A651 Mill 41.2 71.4 29.5 44 44 62 50 Flange 2 A36 68A651 Lab 40.4 72.8 30.5 49 63 25 46 Flange 4 A36 71E820 Mill 43.8 73.5 27.0 54 61 54 56 Flange 4 A36 71E820 Lab 41.3 71.4 31.5 48 52 41 47 Web 1/2 A36 758173 Mill 41.8 69.1 24.0 43 32 43 39 Specified Min A36 - - 36 58*** 23 10 10 10 15

Notes: (1) Tensile properties are average values from two tests. Unless otherwise noted, the yield strengths were determined by tho 0.2* offset and 0.5% elongation mcthods in the laboratory and miii tests, respectively. The specified Charpy V-no-.ch values are for AASHTO Zone 2. * Retest lksi = 6.895 MP ** 0.5% elongation °C = 5/9 (°F - 32) Maximum value of 80 ksi is also specified. 1 in. = 25.4 mis - 1 ft-lb = 1.356 N-rn TABLE 9 TA3LE 10 CHEMICAL COMPOSITION OF PLATES (MILL LADLE ANALYSIS) CHARPY V-NOTCH DATA ON SPECIMENS FROM EXCESS LENGTH OF WELDS Thickness, Heat Element, percan_ Conponent inches Steel No. C Mn P S Si Cu Ii Cr Mo Al V Energy Absorption at OF, ft-lbs Ae 0.04 Flange 2 A586 74B120 0.16 1.13 0.009 3.022 0.28 0.35 (.17 0.49 0.02 0.04 Detail Location I of iii Flange 4 A540 72E579 0.17 :.i 0.011 3.015 0.25 0.35 0.19 0.64 0.02 0.05 0.06 Web 1/2 A589 659225 0.13 1.04 0.009 7.076 0.23 0.31 0.03 0.53 0.01 0.05 0.04 5888 Bottom IStart) 15 15 Specified Max (or Range) ASSB-75 - 0.10 0.97 0.04 3.05 0.15 0.25 3.40 0.02 5889 Top (Finish) 26 31 588C (Grade A) to to to to to to Bottom )Start 10 24 588C Top (Finish) 18 15 19.2 0.19 .25 0.30 0.40 0.65 0.10 36135 Bottom (Start) 23 V. Flange 2 A36 68A651 0.21 0.97 0.009 0.024 0.23 36B5 Top )Finish) 49 ill Specified Max (or Rcnge( 2 A36 - 0.26 0.90 0.04 0.05 0.15 36BS Bottom (Start) 29 23 to to 368S Top (Finish) 13 16 19.5 1.20 0.30 Notes, Specimens cut from Standard location a qiuterthiciress Flange 4 A30 'fH20 3.22 393 0.014 0.029 0.22 0.02 (.17 0.07 0.04 0.047 (with respect to 2-inch plate) on side awiy from tapered face. Specified Max (or lUng..) 4 A36 - 0.27 3.85 0.04 0.05 0.15 tC to 'C = 5/9 ('F - 32) 1in. = 25.4 mm (.71) 0.30 1 ft-lb 1.356 N-rn Web 1/ 2 A36 759173 0.23 1.C1 0.005 0.020 0.018 Specified Max (or Range) 1/2 A36-75 - 0.25 0.04 0.05

1 in. = 25.4 mm

Figure 4. Test SCi uj (looking no, thwest). ('I 16

TABLE 11 TEST RESULTS FOR WELDING-PROCEDURE QUALIFICATION SPECIMENS OF AN ELECTROSLAG 2- TO 4-IN. TRANSITION JOINT WITH A36 PLAIkS Charpy-V-Notch Specimens:

Energy_Absorbed,_ft-lb - - Specimen For 0°F Test Temperature For 400F Test Temperature F- Room Temperature (70, 75°F) Location Individual Values Mean Individual Values Mean Tndividual Values Mean

2' Base Plate* 27, 21, 15 21.0 49, 63, 25 45.7 78, 77 77.5 4" Base Plate* 25, 33, 32 30.0 48, 52, 41 47.0 73, 71 72.0 Weld Metal 17, 23, 12, 11, 10 13.3" 23, 27, 27, 31, 29 27.7"" 45, 57, 51, 41 48.5 HAZ 2" Plate 46, 22, 50, 29, 23 32.75* 52, 74, 64, 52, 43 57.0" 95, 75, 99 89.7 RAE 4" Plate 25, 22, 34, 30, 34 29.0" 36, 38, 28, 46, 31 35.8"" 62, 52, 54 56.0

Tension Specimens:

Tensile Yield Strength,* Strength, Reduction of Elongation In ksi kai - _Area, % 2 In., 8 - Specimen Description and Location - Ind. Val. Mean Ind. Val. Mean Ind. Val. Mean Ind. Val. Mean

Transverse Plate Type, Across Weld - - 71.8, 71.8 71.8 - - - - Round, All-Weld-Metal, Parallel to Weld 49.9, 50.6 50.2 78.6, 79.1 78.8 61.8, 64.8 63.2 27.5, 31.0 29.2 Round, 2-Inch-Thick Plate, at 1/4 Thickness 40.9, 39.8 40.4 72.6, 73.1 72.8 62.5, 63.2 62.9 30.5, 30.5 30.5 Round,+$ 4-Inch-Thick Plate, at 1/4 Thickness 41.7, 40.9 41.3 71.5, 71.3 71.4 62.3, 61.3 61.8 32.0, 31.0 31.5

Side-Bend-Test Specimens (3/4-Inch Bend Radius): All Seven Specimens Passed

Macrospecimens: Passed --

Additional Tests at Different Temperatures 1 ksi = 6.895 MPa Highest and Lowest Values Discarded "c = 5/9 ("F - 32) + 0.2% Offset 1 in. = 25.4 mm ++ Longitudinal 1/2 Inch Diameter 1 ft-lb = 1.356 N-m

TABLE 12 TEST RESULTS FOR WELDING-PROCEDURE QUALIFICATION SPECIMENS OF AN ELECTROSLAG 2- TO 4-IN. TRANSITION JOINT WITH A588 PLATES Charpy-V-Notch Specimens,

Energy Absorbed, ft-lb Specimen For o°F Test Temperature For 40°? Test Temperature For Room Temperature (70, 750F) Location Individual Values Mean Individual Values Mean Individual Values Mean

2" Base Plate" 42, 47, 13 34.0 48, 88, 70 68.7 101, 83 92.0 4" Base Plate" 24, 5, 17 15.3 11, 31 20.0 24, 11 17.5 Weld Metal 10, 17, 11, 17, 12 13.3" 48, 35, 17, 17, 37 29.7"" 40, 41, 29 36.7 HAl 2" Plate 12, 15, 32, 15, 24 18.0"" 22, 51, 37, 33 30.7"" 75, 45, 52 57.3 HAl 4 Plate 17, 9, 13, 10, 10 11.0** 25, 9, 27, 15, 23" 21.05* 45, 21, 43 36.3

Tension Specimens:

Tensile Yield Strength, Strength Reduction of Elongatic'n In ksi k.i Area, 8 2 Zn., 8 Specimen Description and Location Ind. V.1. Mean Ind. V.1, Mean Ind. V.1. Mean Ind. Val. Hear

Transverse Plate Type, Across Weld - - 83.2, 83.2 83.2 - - - - Round,4 All-Weld-Metal, Parallel to Weld 60.5, 59.7 60.1 87.6, 87.9 87.9 55.7, 49.5 52.6 22.5, 19.5 21.0 Jlound,++ 2-Inch-Thick Plate, at 1/4 Thickness 51.9, 52.6 52.2 82.3, 81.2 81.8 71.4, 72.5 72.0 28.0, 29.5 28.8 Round,44 4-Inch-Thick Plate, at 1/4 Thickness 54.7, 57.3 56.0 87.6, 90.4 89.0 67.8, 6.3 66.6 25.5, 25.0 25.3

Side-Bend-Test Specimens (3/4-Inch Bend Radius)t All Se-van Specimens Passed

Macrospecimena Passed - - Additional Tests at Different Teatper.tures 1 ksi = 6.895 MFa Higheat and Leat Values Discarded "C 5/9 ('F - 32) + 0.2% Offset 1 in. = 25.4 mm 4+ Longitudinal 1/2 Inch Diameter 1 ft-lb = 1.356 N-s 17

TABLE 13 MINIMUM ALLOWABLE VALUES FOR WELDING-PROCEDURE QUALIFICATION TESTS

Minimum Value* Test Property Location A588

Charpy V-notch Energy absorbed at 40°F 2-inch plate and HAZ 15 ft-lbs 15 ft-lbs

Charpy V-notch Energy absorbed at 40°F 4-inch plate and HAZ 15 ft-lbs 20 ft-lbs

Charpy V-notch Energy absorbed at 0°F Weld metal 15 ft-lbs 25 ft-lbs

Tension test Yield strength 2-inch and 4-inch plate 36 ksi 50 ksi

Tension test Tensile strength 2-inch and 4-inch plate 58 ksi** 70 ksi

Tension test Elongation in 2 inches 2-inch and 4-inch plate 23% 21%

Tension test Yie]d strength Weld metal 36 ki 50 ksi

Tension test Tensile strength Weld metal 60 ksi 70 ksi

Tension test Elongation in 2 inches Weld metal 24 21% * The'minimum CVN values are mean values; individual values may not he loss than 2/3 of these values. The minimum values are for AASHTO Zone 2. 1 ksi 6.85 MPa ** A maximum value of 80 ksi is also specified, °c 5/9 (F 32 1 in, 2'1.4 mm 1 ft-la 1.356 N-; included in Figures 1 and 2. Figure 4 shows the test were obtained by projecting the stresses measured at gage setup. Loads were applied and controlled by a closed-loop locations to the toes. For convenience, loads rather than cicctrohydraulic testing system (9, 10). Loads were mea- stresses were controlled during each test. The relationship sured with load cells, stresses were measured with electric- between the nominal stress and the corresponding loads resistance strain gages, and temperatures were measured was established by a static calibration that followed proce- with thermocouples. For the low-temperature impulse test, dures used in earlier tests (10, 11). During the tests, an insulated box was placed around the test section and there were three independent checks on the primary load- liquid nitrogen was added to cool this section. The nominal bending stresses at the toe of the tapered control system to assure that the loading was close to the joints were the main test parameters. These stresses, desired loading: (1) strain-gage recordings, (2) limit which excluded residual stresses and stress concentrations, switch settings, and (3) control settings.

ChAPTER TWO

FINDINGS

abnormality noted was the chemical composition of elec- trode C (Table 5). Weld deposits made with this EM 13K- Plate Materials EW electrode contained exceptionally high levels of arsenic; All plates used in this program are considered typical therefore, the as-received wire was analyzed for arsenic, of those currently produced for bridge steels, and, as can antimony, and tin. The arsenic level of 0.044 percent is be seen from Tables 2, 3, and 4, they met the appropriate considered much higher than typical and probably ad- ASTM specification (A36 and A588 Grade A) and the versely affected the notch toughness of the welds. AASHTO specification for Zone 2 service. Nondestructive Inspection Electrodes, Fluxes, and Consumable Guide Tubes In general, the weidments were sound, as determined by The welding consumables used in the program were nondestructive examination. It should be noted that each obtained commercially from various suppliers. The only of the inspection techniques—visual, ultrasonic, and radio- 18

graphic—detect imperfections that are not necessarily nificantly narrower than those of the 4-in.-thick weldments. detected by other techniques. This illustrates the advantage As expected, the size of the HAZ, and thus the peak of using multiple nondestructive-testing techniques for temperatures, was greatest at the top of the weldments. The critical welds. Nearly all the irregularities were associated variation in peak temperature, as evidenced by the heat- with restarts, premature weld stops, or occasional regions tint patterns, is shown in Figure 6 for 1-in.-thick weldments of lack of fusion at tfrie plate surface; thus, the abnormal made with water-cooled shoes and in Figure 7 for weld- areas were visible and easily avoided in specimen prepara- ments made with solid shoes. There was some variation in tion. As will be discussed later, occasional minor grain- thermal distribution from side-to-side, from front-to-back, boundary separations were observed during destructive or along the length of a joint due to such factors as examination of several weldments that were not detected electrode position and shoe contact. In weldments made by prior radiographic or ultrasonic inspection. with water-cooled shoes, peak temperatures at the mid- height were often lower than those at the bottom (Fig. 6). Thermal Distribution However, in weldments made with solid shoes, peak The thermal cycles during and after welding can have temperatures gradually increased with increasing distance a significant effect on the mechanical properties of the along the joint (Fig. 7). weldments. Therefore, detailed studies were conducted, Peak temperatures did not vary with weldment length which included heat-tint patterns for all weldments, and when the weldments were made with a single guide tube thermal cycles were determined by using thermocouples on and water-cooled shoes. selected weldments. The results of these studies may be The use of two fixed-guide tubes versus a single summarized as follows: oscillating guide tube in 4-in.-thick weldments generally resulted in higher peak temperatures. The use of 24-in.-wide (610-mm) plates in this program was more than adequate to simulate thermal The use of solid versus water-cooled shoes resulted in conditions in full-size bridge girders, and the use of plates higher peak temperatures in 1-in.-thick weldments but not as narrow as 12 in. (305 mm) would probably be sufficient in 4-in.-thick weldments. However, the solid shoes heated for procedure qualification testing. Typical thermal cycles considerably during welding, which slowed the weld-metal at various distances from the groove face in a 4-in.-thick and HAZ cooling rates. weldment are shown in Figure 5. Note that peak tempera- The cooling time in the austenite-to-ferrite trans- tures near the weld had decreased to temperatures below formation-temperature range 800 to 500 C (1470 to 930 F) those at which most metallurgical reactions would occur was generally shorter at the midheight rather than at the top (for example, 500 C or 932 F) before the temperature at or bottom of the joint, was shorter for water-cooled versus a distance of 12 in. had risen significantly. solid shoes, and was shorter for single oscillating versus The HAZ's of the 1-in.-thick weldments were sig- two fixed-guide tubes.

1200

NUMBERS SHOWN ARE DISTANCES I., FROM GROOVE FACE '°°° I I L

800- I 1500 p. \ Ifb..11/2 li\ I J END OF SHOES Cc WELD REMOVED 60O- Iji k% 1000 I— I-

4 400- II, /11 1 inch 25.4 ,nn, - 500 200

—_ 18 24 C I i I I I 0 20 40 60 80 100 120 140 160 TIME. minoten Figure 5. Thermal history at midheight of 4-in.-thick electroslag weldment (weldment No. 8). Figure 6. Heat-tint pattern on surface of 1-in.-thick elect roslag weldnent (No. 17) made ivith tt'ater-cooled s/toes (X 1 / 10).

Macroscopic Examination austenite grain size had no relationship to the cell structure. Thus, the weld metal probably solidified as delta ferrite Figure 8 shows a typical macrostructure of an ESW and subsequently transformed to austenite of various region in three orientations. Table 14 summarizes certain grain sizes. geometric features of the weld bead for all the transverse As shown in Figure 8, the fused region is revealed by a sections. The admixture, defined as the ratio of fused base mixture of coarse and fine prior-austenite grains that metal to total fused metal, was calculated by determining originate at the base-metal—weld-metal interface (Fig. the groove cross-sectional area and measuring the final SB), and at the shoe interface (Fig. SC), and grow in- fused area. The fine-grain area associated with the center wards and upwards. of the fused area was often well definci in the A588 weldments and could be measured. The percentage of fine- Residual Strains grain area in the other weldments was estimated visually. Electroslag weldments differ considerably from struc- The grain size referred to in this section is the prior- tural arc-welded butt joints in that they are made in one austenite grain size, which is readily revealed by macro- pass and have very high heat input. These differences are examination and microexamination by using conventional expected to have a major effect on the resultant residual- etchants. However, as discussed in the section on micro- stress pattern which, in turn, could affect the fatigue scopic examination, special etchants that reveal the solidi- performance and propensity to cracking; therefore, the fication structure showed that the weld metal consisted of residual-stress pattern of selected weldments was studied a uniform, fine cellular structure, and that the prior- in detail. This study is summarized as follows: 20

Figure 7. Heat-tin: pastern on surface of 1-in.-thjck electroslag weld,nent (No. 13) ,nade with solid shoes (X i/JO).

The shrinkage across the weld, after welding, included Elements such as arsenic are not normally encountered a fixed contraction (which varied with plate thickness) in filler metals. When present, they generally have an at the point where the welding started and an additional adverse effect on notch toughness and are particularly shrinkage that was linear with weldment length (Fig. 9). detrimental to weldmcnts in elevated-temperature service. Residual through-thickness strains, measured in a This particular electrode was obtained during a period of 4-in-thick weldment, showed that the residual through- severe filler-metal shortages, and the arsenic level is not thickness stresses were compressive (Fig. 10), with a typical of this type of electrode. magnitude less than the magnitude of strain at the uniaxial yield stress (1700 microin./in.). Microscopic Examination The center thickness portions of the weidments con- tained residual tensile transverse strains and compressive The niicrostructure of each type of weldment was longitudinal strains. The compressive residual strain was examined in detail. Figure II shows typical microstruc- less than about one-half the magnitude of strain at the tures of weld metals. All the weld metals contained ,in yield stress. intermittent-to-continuous network of proeutectoid ferrite No significant differences in residual stresses were outlining the prior-austenite grain boundaries and Wid- observed for different wire-feed techniques. manstätten-ferrite side plates extending into the matrix. In general, the magnitude of the residual stresses found Acicular ferrite was usually the major component within in this study was much less than originally anticipated. the matrix. The weld metal usually exhibited much larger Additional work is required to explain the through- austenite grains in the 4-in-thick veldments than in the thickness compressive residual-strain pattern observed in 1-in-thick weidments, and the procutectoid ferrite on the the 4-in-thick weldment. prior-austenitc grain boundaries was usually outlined by pearlite. The 1-in-thick weidments exhibited a mixture Chemical Composition of Weld Metal of coarse and fine austenite grain sizes in the weld metal, with discontinuous networks of procutectoid ferrite and The chemical composition of the deposited weld metals little or no pearlite. The weidments made with the solid is given in Table 15. Analyses of the residual elements copper shoes contained a significant amount of polygonal (boron, columbium, arsenic, antimony, and tin) were ferrite and pearlite within the matrix. conducted on either a representative weld metal or the filler The solidification structure of the weld metal was metal. Because filler metal C (EMI3K-EW) was found to revealed by etching with a saturated aqueous solution of contain a significant amount of arsenic (0.044%), several picric acid with a vetting agent. The structure appeared to of the weld metals deposited with this filler metal were also be cellular, with cell size of the same order of magnitude as analyzed for arsenic. that of the fine austenite grains. However, the cell size was 21

A. Transverse section *

• H. v : : ix .!'j;• pk

- A B. Section parallel H to plate at mid- thickness. -; , e• .. ;- - 1 it!

C. Section normal to plate at midwidth of weld.

Figure 8. Macrost,-ucture of weldment (No. 8) 4-in-thick A588 with oscillating consumable guide tube (,nixed acids etc1zant X 1). 22

TABLE 14 GEOMETRY OF WELD METAL IN TRANSVERSE SECTIONS Plate Fine Grain Weldment Thickness, Specimen Fused Admixture, Area, No. inches Location Area, in.2 inch2 Percent

1 1 Top 2.89 53 _* Mixed Mid 1.92 29 - 0 Bot 2.14 36 -

2 1 Mid 1.75 22 - Mixed

3 4 Top 13.44 65 0 0 Mid 8.30 43 0 0 Bot 8.05 41 0 0

4 4 Mid 10.28 54 0 0

5 4 Mid 4.83 2 - Mixed

6 1 Top 2.65 49 - 80 Mid 2.07 34 - 85 Bot 2.33 42 - 90

6A 1 Top 2.68 49 - 90 Bot 2.16 37 85

7 1 Mid 1.73 21 - 90

8 4 Mid 9.27 51 1.32 14

8A 4 Top 11.06 57 6.57 59 Bot 8.82 46 1.99 23

9 4 Top 17.65 73 4.54 26 Mid 12.17 61 3.15 26 Bot 10.72 56 4.20 39

10 4 Mid 4.99 5 0 0

11 1-2 Mid 2.90 40 1.44 50

12 2-4 Mid 4.93 42 0.84 17

13 1 Top 2,64 49 - Mixed Mid 2.45 45 - Bot 2.28 40 -

14 1 Mid 1.92 29 - 80 15 4 Top 13.41 65 4.19 31 Mid 10.09 53 1.11 11 Bot 8.71 46 1.41 16

16 4 Mid 10.80 56 1.38 13

17 1 Mid 1.98 31 - Mixed

18 4 Mid 8.37 43 0 0 & 19 1 Mid 1.87 27 - 75

20 4 Mid 11.08 57 0.08 0.7 * Could not be measured because of mixed distribution. Conversion Factors:

1 inch 25.4 mm 1 square inch = 645 mm2

23

45 0 I I I I • I I I

WELOMENT NO. 12 to 0 0)2" AND 4" THICKNESS) 0 40 ØWELDMENT NO. 11 0 0" AND 2' THICKNESS) 0 0 OAFTER SAW CUTTING RESIDUAL DAFTER GRINDING TO DO COMPRESSION 5/8-INCH WIDTH QAFTER 1/4-INCH SLICING. 35 1 inch 25.4 mn, 2 0 AFTER FINAL MACHIN- 0 0 ING 0 11 DO A. 30 - 3 0 0 LOCATION B Co 0 0 z 0 U 00 I I I I I I I I I- +3000 +2000 +1000 0 -1000 -2000 -3000 I 0 0 I, 0 I I I I I I I- A3 0 0 o 2 B. 0 LOCATION D O 0 at 0 4 0 U 15 DO 0 RESIDUAL 04 INCHOF COMPRESSION D 0 HRINKAGE PER /INC. OF WELD LENGTH 2 A0 00 10 00 0

3

5 00 - 0 BOTTOM OF WELD dD I I I I I +3000 +2000 +1000 0 -1000 -2000 -3000 001 0 I I STRAIN RELIEVED, 1a6 inch/inch 0.0 0.15 0.10 0.15 0.20 0.25 0.30 AVERAGE SHRINKAGE ACROSS WELD. inch Figure 10. Through-thickness strains normal to plate surface, Figure 9. Average shrinkage across the weld 0/transition relieved after removal of varying amounts of material weldments (No. 11, No. 12). (.weldment No. 16).

uniform throughout the weld and was not associated with atmosphere 0.25-in, thick specimens cut from the surfaces either the coarse or fine prior-austenite grains as delineated of the welded plates. The specimens were exposed at by the proeutectoid ferrite. This suggests that the weld Monroeville, Pa., using the standard ASTM orientation metal solidifies as delta ferrite and subsequently transforms (30 degrees from the horizontal, facing south) and ex- to austenite, which then transforms to proeutectoid ferrite amined periodically. At each inspection, the colors of all and other transformation products. Grain-boundary separa- welds and the base metal were identically matched. At the tions were observed in the weld metal in weldments 1, 3, last inspection, after 31 months of exposure, the specimen and 17. exhibited on both the weld surfaces and the base metal the The microstructures of the HAZ in the 1- and 4-in.-thick tightly adherent red/blue oxide typical of mature ASTM A36 and A588 steel weldments are shown in Figure 12. All A588 Grade A steel. the steels exhibited significant grain coarsening in the high-peak-temperature region, with proeutectoid ferrite Mechanical-Property Tests outlining the prior-austenite grain boundaries. The matrix in the HAZ in the 1-in.-thick A36 steel (Fig. 12A) con- Tension Tests sisted principally of bainite with some pearlite adjacent to The results of duplicate longitudinal 0.505-in.-diameter the proeutectoid ferrite. In the HAZ of the 4-in.-thick A36 (12.8-mm) all-weld-metal tension tests are given in Table steel (Fig. 12B), pearlite was the major constituent, with 16. The results of duplicate, transverse plate-type tension a significant amount of Widmanstätten-ferrite side plates. tests across the weldments are presented in Table 17. As In HAZ of the A588 steel, bainite was the major phase in expected, there was no problem in meeting the minimum the 1-in.-thick weldment (Fig. 12C). In the 4-in.-thick tensile-property requirements of the base plate in all weld- weldment of A588 steel (Fig. 12D), the major constituents ments with the filler metals selected. in the matrix were bainite and ferrite side plates with As to the location for testing of the all-weld-metal regions of blocky pearlite. There was little effect of shoe tension-test specimen, there was no consistent variation type (solid versus water-cooled) on the microstrutures from surface, to quarter point, to midthickness. For proce- observed. dure qualification testing, the data generated in this study indicate that the current procedure of centering the speci- Weathering Tests men at the quarter point is satisfactory. The color match between electroslag welds and A588 The significant result noted in the all-weld-metal tensile Grade A base metal was determined by exposing to the data was the lower yield and tensile strengths of welds TABLE 15 CHEMICAL COMPOSITION OF ELECTROSLAG WELDS-PERCENT

We1ent No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 39 19 20

Steel Grade: A36 A36 A36 A36 A36 A588 A588 A588 A588 A588 A588 A588 A36 A588 A588 A588 A36 A36 A588 A588

Thickness, 1 1 4 4 4 1 1. 4 4 4 1-2 2-4 1 1 4 4 1 4 1 4 inches Electrode: A A A A A C C C C C C C A C C C B B 0 D

C 0.13 0.09 0.17 0.18 0.07 0.13 0.12 0.15 0.17 0.13 0.12 0.13 0.11 0.12 0.14. 0.15 0.16 0.19 0.13 0.14 Mn 1.29 1.42 1.36 1.25 1.47 1.19 1.28 1.17 1.12 1.15 1.08 1.07 1.34 1.09 1.12 1.15 1.30 1.29 0.79 0.81 P 0.011 0.011 0.007 0.008 0.010 0.017 0.017 0.015 0.013 0.014 0.017 0.017 0.010 0.017 0.014 0.014 0.009 0.008 0.009 0.009 S 0.022 0.022 0.021 0.025 0.019 0.027 0.022 0.019 0.019 0.016 0.022 0.020 0.022 0.023 0.018 0.018 0.024 0.023 0.022 0.015 Si 0.21 0.24 0.32 0.28 0.25 0.39 0.48 0.36 0.31 0.43 0.35 0.35 0.25 0.43 0.36 0.35 0.11 0.18 0.30 0.23 Cu 0.03 0.03 0.03 0.05 0.04 0.14 0.11 0.19 0.20 0.07 0.15 0.16 0.03 0.10 0.20 0.19 0.10 0.10 0.35 0.39 Ni 0.03 0,03 0.03 0.03 0.03 0.04 0.04 0.16 0.18 0.06 0.06 0.12 0.03 0.04 0.16 0.16 0.04 0.07 0.29 0.39 Cr 0.04 0.04 0.03 0.04 0.02 0.25 0.18 0.28 0.36 0.11 0.27 0.24 0.03 0.14 0.27 0.28 0.05 0.08 0.54 0.52 Mo 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.03 0.04 0.01 0.01 0.02 0.01 <0.01 0.03 0.03 0.01 0.02 0.02 0.04 V <0.005 - <0.003 - - 0.02 0.015 0.03 0.04 0.01 0.02 0.02 - 0.01 0.03 0.03 <0.005 0.01 0.02 0.03 Ti <0.005 - 0.001 - - <0.005 - <0.005 <0.005 - <0.005 <0.005 - 0.005 0.005 0.005 <0.005 <0.005 <0.005 <0.005 Al(total) <0.002 <0.002 0.009 0.005 <0.002 0.006 0.004 0.007 0.008 0.002 0.006 0.006 <0.002 0.005 0.008 0.007 <0.002 0.008 0.007 0.007 N 0.003 0.003 0.005 0.007 0.002 0.006 0.006 0.005 0.005 0.005 0.006 0.006 0.003 0.006 0.005 0.006 0.005 0.005 0.006 0,006 B - - 0.0001 - - - - 0.0002 - - - - - - - - - - - - Cb - - <0.004 - - - - <0.005 - - - - - - - - - - - - As <0.010 - <0.010 - - - - 0.024 - - 0.028 0.030 - 0.040 0025 0.022 <0.01 - <0.01 <0.01 Sb <0.010 - <0.010 - - - - <0.010 - - - - - - - - - - - Sn <0.002 - 0.004 - - - - 0.003 - - - - - - - - - - - - 0 0.0241 0.0298 0.02190.0207 0.0190 0.02450.0335 0.0220 0.02070.0223 0.0201 0.0235 0.0257 0.02190.0229 0.0267 0.0222 0.0207 0.0198 0.0201

Conversion Factor:

1 inch = 25.4 nun 25

made with solid copper shoes when compared with the nounced at low values of the stress-intensity-factor railge strengths of similar welds made with water-cooled copper (iNK). shoes (see Table 18). It is assumed that the slower cooling The retardation effect was substantially greater for rate associated with the use of solid shoes is responsible. crack propagation in the coarse-grained-HAZ and bond-line These data were transmitted to the AWS Structural Weld- regions than in the weld metal, but this was not consistently ing Committee for their consideration because the current observed. Structural Welding Code (AWS 131.1-75) listed changes in The variations in the microstructures of the weld welding shoes as a variable that requires only radiographic metal and the HAZ caused marked differenc&s in the or ultrasonic testing for requalification. They now require topography of the fatigue-crack surfaces, but had no mechanical-property testing for requalification in the event significant influence on the crack-propagation rate. of change of shoe type. Evaluation of the crack-growth behavior of the weld- Three of the six tension tests of' weidment No. 3 and ments in terms of a cyclic-life parameter, N1, gave results one of the two tension tests of weidment No. 17 showed consistent with those based on incremental crack-growth abnormally low elongation and reduction-of-area values, data. although the yield- and tensile-strength values were typical The observed retardation in crack-growth rate in the (Table 16). Examination of these specimens suggested that weldments is attributed to compressive residual stresses this behavior was caused by grain-boundary separations. introduced by welding; this effect was analyzed in terms of a stress-intensity-factor-suppression concept. (Compressive Hardness in the Weld residual stresses retard crack growth by crack closure effect; tensile residual stresses do not have such effect and, there- The results of typical Rockwell B hardness traverses fore, would not appreciably effect growth rate.) Because across transverse sections of the weldments are shown in of the inevitable redistribution of residual stresses in the Figures 13 and 14. The hardness values were considerably process of obtaining the fatigue-test specimens from the more uniform than those often experienced in arc-welds. whole weldments, the residual-stress pattern in the speci- The hardest portion of the A36 steel welds usually was the mens tested was probably different from that in the weld- weld metal, with the hardness of the HAZ intermediate ments as a whole. Nevertheless, based on the observation between that of the weld metal and the base metal. Unlike that tensile residual stresses have no appreciable effect on the A36 weldments, the HAZ's of the A588 weldments crack-propagation rate (under tension-to-tension loading were slightly harder than either the base metal or the weld conditions), whereas compressive residual stresses have a metal, both of which were similar in hardness. This is to retarding influence, the general conclusion that fatigue- be expected because of the greater hardenability of this crack-growth rate in properly made electroslag weldments steel. is similar to or slower than the rate in the base steels All hardness values were well below the level at which would be just as applicable to whole weldments. stress-corrosion cracking could be anticipated. A conservative estimate of the fatigue-crack-propaga- tion rate in properly made electroslag weldments of low- Fatigue-Crack-Growth Rate strength structural steels may be made by using the upper bound of the scatter band (with respect to crack-growth Fatigue behavior and notch toughness are the two major rate) of the crack-growthdata for the base steels used in concerns about the use of electroslag weldments in bridges. the present investigation, as given by the relationship Fatigue-crack growth-rate studies were conducted on six 1-in.-thick and four 4-in.-thick representative butt-welds da 1.68 10- (K)33 (1) in the as-deposited condition. Various orientations and x locations of the fatigue crack in relation to the weld geom- where da/dN is the crack-growth rate in inch per cycle and etry were investigated (see Fig. 15). The specimens were inch. notched at the weld centerline and positioned at mid- K is the applied-stress-intensity-factor range in ksi V On this basis, the prediction of the service life of thickness. The crack-growth data obtained were analyzed by using linear-elastic fracture mechanics. The main 1nd- electroslag-welded structures subjected to cyclic loads and ings are summarized as follows: the determination of safe inspection intervals may be made by the same approach as that presently used for standard Considerable resistance to crack propagation was structural-steel weldments made by other. processes. encountered during precracking (initial crack extension before obtaining craek-growth data) of the fatigue-test Fracture-Toughness Tests specimens, indicating that properly made electroslag weld- ments would generally have fatigue lives equal to or.longer Along with fatigue performance, resistance to brittle than those of structural bridge steels for small initial crack fracture is a major concern of engineers and designers on size. the use of electroslag weldments in bridges. Historically, Fatigue-crack-growth rates, da/dN, in the weldments CVN tests have been used in assessing the notch toughness were similar to, or up to five times slower than, the rate in of welds, and these tests were extensively used in this the base steels (Fig. 16). The da/dN data were obtained investigation. Concern was expressed, however, that the at tK values above threshold range (Region 1). CVN specimen was too small for use on electroslag welds The retardation in crack-growth rate was most pro- in that the area tested is the same order of magnitude as 26

-

;J4. --.-'•;:-.- /

.CL7S V.

A. Filler metal A in B. Filler metal A in 1-inch-thick weldment 1. 4-inch-thick weldment 3.

C. Filler metal C in D. Filler metal C in 1-inch-thick weldment 6. 4-inch-thick weldment 8. 1 inch = 25.4 mm

Figure 11. Typical ?nicrostructures of weld tuetals deposited with electrodes A and C by using water-cooled shoes (nilal and picral c/chant, X 100).

the grain size. Conceivably, the specimen could be tested formance to correlate with toughness as measured by within a grain or a grain boundary and not the over-all specimens more amenable to control and procedure composite. For this reason, other fracture-toughness tests qualification testing. (drop-weight and dynamic-tear tests) were also included in the program as possible control tests for acceptance C/Earp)' V-Notch Impact Tests criteria. Finally. K1, determinations were attempted in an Over 1500 CVN tests were conducted and individual effort to establish a base line for required service per- test results for the weld metal and the HAZ were included

27

'-S - •- -. . .-.1 P' , .. -:... . \•'-; .+.l 4.• . rt ••' ..

I

A. One—inch—thick A36. B. Four—inch—thick A36.

C. One—inch—thick A588. D. Four—inch—thick A588. 1 inch = 25.4 mm Figure 12. Typical ,njcrostructures of heat-a/Jec:ed-zones in A36 and A588 steel weldments (nunl and picral etcizant, X 100).

in the Phase I final report. Individual CVN values for the Analysis of the CVN data for the weld metal failed to weld metal showed considerable scatter. This is shown show a consistent variation along the weld length; however, graphically in Figures 17 and 18 for 1- and 4-in.-thiek in some instances the notch toughness at the extreme top weldments, respectively. These results encompass all the and bottom of the weld was inferior and not representative previously mentioned test locations within the weld metal. of over-all properties. The results for tests conducted at 0 F (-18 C) are sum- Also, weld length did not affect the notch toughness. marized in Table 19. These results show a considerable Therefore, for acceptance testing, the location of specimens range of values depending on steel grade, filler metal, and as specified in AWS Dl.! -77 appears satisfactory. welding procedure. Because this project was designed There was a significant variation in CVN energy absorp- to develop acceptance criteria, correlation between the tion depending on location of specimens with respect to CVN properties and welding variables was not attempted. the weld centerline. The weld centerline exhibited the tj TABLE 16 TABLE 16 (Continued) 00 Tensile Elongation Reduction ALL-WELD METAL TENSILE PROPERTIES weidment Test Yield Strength Strength, in 2 Inches, of Area. No. Location (0.2% (ffset), psi psi Tensile Elongation Reduction Weldrnent Test Yield Strength Strength, no 2 Inches, of Area, . 10 5 48,000 74,200 29.0 66.8 No. Location (0.2% Offset), psi psi % 9 S 49,700 75,500 25.0 66.8

M 53,500 75,500 25.5 64.7 Q 48,900 74,900 29.0 68.1 N 50,100 74.000 25.5 66.9 Q 47,600 74,500 29.0 69.4

2 .0 40,900 66,800 30.5 69.5 N 46,400 72,300 29.0 67.6 8 42.400 65,800 32.0 72.4 8 48,100 74,200 29.0 65.2

.3 S 49,800 76,700 27.0 60.0 11 8 55,000 81,300 26.5 64.0 S 49,900 76,900 28.5 60.0 8 54,400 80,800 27.0 65.4

- Q 52,100 76,000 12.0 24.8 12 H 58,100 85,200 25.0 64.1 Q 52,700 75,300 11.0 22.3 8 54,000 81,400 27.0 66.1

8 51,000 77,900 25.5 53.5 13 H 44,800 68,900 27.5 69.9 M 54,000 76,500 11.5 21.6 M 47,500 70.300 23.5 51.6

4 S 47,400 76,700 30.5 66.2 14 M 50,700 75,600 28.5 66.1 S 48,600 78,000 29.5 65.6 H 52,400 75,900 27.5 63.3

Q 50,900 78,600 29.0 64.2 . 15 S 51,300 79,600 27.0 55.2 Q 48,400 77,700 30.5 65.1 S 51,300 79,900 27.5 67.0

8 50,000 79,500 28.0 63.5 9 49,400 79,400 27.5 65.2 M 49,800 79,500 28.0 59.7 Q 50,800 80,500 30.0 65.8

5 . 5 37,700 60,300 34.0 75.1 M 49,000 78,600 28.0 67.1 S 38,200 59,800 30.0 74.5 - 8 49,100 79,000 28.5 65'.7

Q 40,400 61,100 32.5 75.3 16 S 57,200 86,000 25.5 65.5 Q 40,100 60,700 32.5 75.4 5 54,000 85,700 28.0 64.9

Q 58,300 87,100 25.0 56.5 M 39,500 60,800 30.5 74.2 57,200 86,700 29.5 65.9 H 39,900 60,300 32.5 74.1 Q

M 55,100 85,300 27.0 65.3 6 H 57,900 83,375 25.0 66.4 - M 56,800 86,300 27.0 65.5 M 56,300 82,957 24.0 66.0 17 H 58,300 81,300 22.0 63.6 7 M 52,200 77,100 28.0 65.8 H 59,900 79,200 12.0. 22.8 H 55,200 80,100 26.5 66.1 18 S 48,900 79,100 29.0 8 5 58,600 85,700 25.0 62.1 63.0 S 46,700 78,400 28.0 62.0 S 58,400 85,400 25.0 60.0 49,400 79,800 29.5 64.7 Q 56,400 84,400 25.5 65.7 Q 4 9,000 78,600 31.0 63.8 9 58,500 86,000 26,5 63.3 Q M 55,100 83,000 26.0 63.5 8 49,500 78,800 25.5 56.5 H 54,100 83,700 25.5 62.4 M 50,900 80,200 20.9 58.0

9 S 59,800 87,600 22.0 58.0 19 H 60,000 87,000 25.0 61.2 58,500 ' 87,100 24.5 56.5 H 59,800 86,300 25.0 55.8

9 60,500 86,500 25.0 60.6 20 S 58,900 84,400 . 20.0 55.0 Q 61,400 85,800 23.0 00. S 53,700 83,400 21.0 43.4

8 58,200 84,800 25.5 Q 58,600 84,500 27.0 58.4 8 56,300 8-.,300 28.0 64.7 9 58,500 84,600 24.0 60.9 Conversion Factors M 56,400 83,100 23.0 55.8 1 psi 6.89 . H 54,900 83,790 24.0 64,3 1 inch 25.4 n

29

lowest impact properties, with properties increasing as the TABLE 17 specimen location was moved toward the weld surface or TRANSVERSE TENSILE PROPERTIES OF toward the bond line. The fine-grain area often observed ELECTROSLAG WELDMENTS * at the weld center had CVN properties inferior to those of Plate Weidment Steel Thickness, Tensile Location specimens located in the coarse prior-austenite grains. NO. Grade inches Strength, ksi of Fracture However, even when the weld center was coarse grained, the center properties were inferior. Figure 19 shows the A36 1 68.4 B 2 A36 1 70.1 N average CVN energy-absorption values as a function of 3 A36 4 63.5 N 4 A36 4 75.9 B test location for all 4-in.-thick weidments; these data were 5 A36 4 68.9 N for the midlength position. As will be discussed later, the 6 A588 1 83.9 N 6A A588 - 1 85.5 W quarter-thickness position at the weld centerline is a prac- 7 A588 1 79.6 B 8 A588 4 83.8 N tical point for specimen location and could be considered BA A588 4 77.2 B representative of the average toughness. 9 A588 4 77.1 B 10 A588 4 76.4 S The HAZ properties also showed considerable scatter. 11 A588 1-2 77.0 B Figures 20 and 21, respectively, show typical plots of 12 A588 2-4 77.8 Ii 13 A36 1 68.8 13 energy absorption vs temperature for 1- and 4-in.-thick 14 A588 1 79.0 N 15 A588 4 81.6 N weidments. Individual values were included in the Phase I 16 A588 4 81.4 N report, and the results at 0 and 40 F (-18 and 5 C) are 17 A36 1 72.6 B .18 A36 4 74.9 B averaged and presented in Table 20. The major significance 19 A588 1 79.2 8 of these data is that some of the values are sufficiently low 20 A588 4 84.2 5 Averaged duplicate plate-type specimens. that HAZ toughness should be considered in qualification * B - base metal testing of electroslag weldments for bridges. As to trends, N - weld metal Conversion Factors: there is evidence that there is a minor degradation of HAZ 1 inch = 25.4 mm impact properties at the top of the weld. This is to be 1 ksi = 6.894 MPa expected because of the previously discussed heat buildup in this region, which causes larger grain size.

TABLE 18 EFFECT OF SHOE TYPE ON TENSILE PROPERTIES OF ALL-WELD-METAL DEPOSITS (AVG. OF DUPLICATE TESTS) - Water-Cooled Solid Copper Shoes Copper Shoes Yield Strength Tensile Yield Tensile Strength, Strength, / Electrode Guide (0.2% Offset), Strength, Steel Type Tubes* Location** ksi ksi ksi ksi

1" A36 A 1 Sta M 51.8 74.8 46.2 69.6

1" A588 C 1 Sta M 57.1 83.2 51.6 75.8

4" A588 C 1 OSc S 58.5 85.6 51.3 79.8 Q 57.5 85.2 50.1 80.0 M 54.6 83.4 49.1 78.8

4 A588 C 2 Sta 5 59.2 87.4 55.6 85.9 Q 61.0 86.2 57.8 96.9 57.3 84.6 56.0 85.8

* Sta - Stationary, Osc - Oscillatinq. Conversion Factors: N - MidthiCknoss, S - Near Surface, Q - Quarterpoint. 1 inch = 25.4 mm 1 ksi = 6.894 MPa

Drop-Weight Tests The results of the drop-weight test NDT determinations The correlation between the drop-weight NDT tempera- are given in Table 21. ture and the CVN results was not very. good. Table 22 These tests showed that the NDT temperatures of both shows the CVN energy absorbed at the NDT temperature. the weld metal and the adjacent HAZ were generally better There does not appear to be a consistent level of energy (lower) than that of the corresponding base metal. Also absorption at the NDT temperature. Lateral expansion the NDT temperatures, as measured by the tests, were not values exhibit a similar relationship to NDT temperature. sensitive to direction of crack propagation (the NDT The relatively low NDT temperature values indicate temperatures, as measured by cracks propagating from the good notch toughness in the weld regions. The NDT surface into the weld, were essentially identical to the NDT temperature test did not distinguish between variations in temperatures for cracks propagating along the weld axis). welding parameters.

30

DISTANCE, cm DISTANCE. cm

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 I 1 .3 4 V C I C V II) II 90 I I I I I I I I I

80 4I

70 TOP TOP

90 IN WELD LOCATION BASE METAL HAZ BOND LINE HIETAL SURFACE A A 4 a LOCATION BASE METAL HAZ BOND LINE WELD METAL .QUARTERTI4 0 0 SURFACE A A 4 a MIOTH 0 0 OUARTER TH D a 0 U MIOTH 0 0 4 •

MIDDLE 70

1%

90 90

80

BOTrOM

70 I 2 3 4 5 6 1 2 3 4 5 DISTANCE. inches DISTANCE. inches Figure 13. Hardness of weldment No. 1. Figure 14. Hardness of weidment No. 6.

Dynamic-Tear Tests K10 Determinations The values for energy and shear as ,a function of test Only one fracture-toughness test at —30 F satisfied the temperature are plotted in Figures 22 through 25. ASTM requirements for a valid K10 test. The one valid These results do not show as great an improvement in K10 test was obtained from the 4-in.-thick A36 steel weld- notch toughness as was observed in the drop-weight test. ment and resulted in a K50 value of 56.7 ksi\/in. The transition temperatures were generally at room temp- (62.3 MPa\/m). The limited fracture-toughness tests erature or above. Except for the performance of the 4-in.- yielded a number of essentially valid fracture-toughness thick A36 plate, which had been normalized, the weld metal (almost satisfying ASTM requirements) values, which were and the HAZ did, however, exhibit better dynamic-tear- approximately equal to or greater than those satisfying the test performance than the base plate, with the toughness of AASHTO fracture-toughness requirements for bridge steels the weld metal being better than that of the HAZ. This having a 50 ksi (345 MPa) minimum yield strength or test is not recommended for production testing because the lower. information obtained does not contribute significantly more information than that obtained by CVN tests. CVN tests Grain-Boundary Separations are much more amenable to production testing under the current state of the art. A discontinuity that has been observed in electroslag

31

p6"

40 SCAER BAND FOR BASE STEELS .1

CRACK ORIENTATION—TL I f .1;'.

!

1 inch = 25.4 mm I ksi,Jinch = 1.1 MPa./rn

_il I I 1 I I 10 20 3040 6080100 CRACK ORIENTATION—TS Figure 15. Orientations and locations of the fatigue crack in a STRESS-INTENSITY-FACTOR RANGE. AK, ksi v'inch 4-in.-thick weldment. Figure 16. Fatigue-crack-propagation data for weldment No.1.

welds is separation at grain boundaries. This is usually Relatively few separations were found in this study; referred to as "microcracking," which is a misnomer however, they do require attention. Similar separations because these crack-like indications can be on a macro- have occasionally been observed in other electroslag weld- scopic scale and, therefore, should properly be referred to ments, but have not been studied in great detail. On the as grain-boundary separation. During the present investi- basis of studies èonducted as part of this project, the fol- gation, grain-boundary separations were detected on lowing observations are noted: metallographic examination of weldments No. 1 and No. 3, in fatigue-crack-growth tests of weldment No. 3 (at com- The separations did not come to the surface. They pletion of test), and in fractured tension-test specimens were usually found from the quarter-thickness point inward from weldments No. 3 and No. 17. A typical grain- in a region of transverse tensile residual stresses. boundary separation (unstrained) in weldment No. 3 is Separations occurred at prior-austenite grain bound- shown in both the unetched and etched conditions in aries that were outlined by proeutectoid ferrite. Figure 26. Note that the separation was confined to the No measurable chemical segregation was detected proeutectoid ferrite outlining the prior-austenite grain on the fracture surfaces in scanning-electron-microscopy boundaries. These separations were generally observed to and photoelectron-spectroscopy studies. be about ½ in. (3 mm) in length on metallographic Metallographic examination with an etchant to reveal examination of transverse sections. the solidification structure showed that the original as-cast In production testing of electroslag weldments, the delta-ferrite grain boundaries generally did not coincide presence of grain-boundary separations is best detected with the observed prior-austenite grain boundaries where by the four side-bend tests required by AWS D1.1. This the separations occurred. This indicates that the separa- code does not permit any discontinuities over ½ in. tions did not occur, during solidification: (3.2 mm) in length after plastic straining (bending). Their appearance and orientation on a macroscopic

32

TEST TEMPERATURE, C TEST TEMPERATURE. C - -10 0 10 20 I . I I 0

80 10 STANDARD AWS TEST LOCATION NON-STANDARD TEST LOCATION 00 80

70 0 100 0 70 0 60 C &90 a 0 0 0 60 50 80 U, C-' C = ° 2- (2 00 , 701 40 50 z 0 z C 00 > >z .( 0 U U 2- 30 (2 0 00 Cz 40 00 0 0 20 U

30 0

10 : 80 0 -30 20 0 -20 0 20 40 60 80 TEST TEMPERATURE, F 10 Figure 17. Individual Charpy V-notch energy absorption values for weld metal in weidment No. 1. 10

I I I I -20 0 20 40 60 Rn TEST TEMPERATURE, F Figure 18. Individual Charpy V-notch energy absorption values for weld metal in weidment No. 3.

scale are very similar to the hydrogen-induced "flakes" that are sometimes found in steel blooms, indicating that hydrogen is a possible cause. Fractographic examination of a fatigue-test specimen from weldment No. 3 revealed that the surface of the / fissure consisted of a cleavage crack with multiple origins. Examination of these crack origins at higher magnification I17.6 169 - revealed the presence of micropores containing small globular inclusions. The inclusions were identified as 164 1166 - manganese sulfide by means of the energy dispersive X-ray analyzer. This is to be expected because cracks will often

4 inchn initiate at the weakest point. The inclusion was very small, (102 mm) and complete elimination of inclusions was not considered 19.3 11.0 I I feasible. The separations did not appear to affect the rate NOTCH LOCATION of fatigue-crack propagation. The separations were not detectable by radiography and were not always detectable by ultrasonic testing in accordance with AWS D1.1-76. High-sensitivity ultrasonic inspection may possibly detect these separations; however, it is very difficult to 1 ft-lb - 1.358 J distinguish them from the ultrasonic reflections from Figure 19. Variation in Charpy V-notch energy absorption with sound, large grain boundaries. test location within a weidment (in ft-lb). 9. Measurements of weld size, degree of admixture, 33

TABLE 19 TABLE 19 (Continued) . Locatron of Tost Spectmen CVII Energy SUMMARY. OF WELD METAL CVN ENERGY Weldment Thickness, Along Along In Absorbed, ABSORBED No. in. Length' Width' Thickness' ft-lb AWS Method—Average at O'F or -18' C) 11 1-2 M 01 0 10 LocatiOn of Test Soecimen CVN Energy Weldinent Thickness, Along Along In Absorbed, 12 2-4 M M Q 8 No. in. Length' Width' Thickness' ft-lb 13 1 B M 01 43 B M M 11 01 M 01 48 01 01 M 21 T M 14 38 T M 01 39 14 1 M M M 10 2 1 M M 01 25 15 4 01 M S 11 3 4 5 01 S 23 01 01 0 8 B M Q 16 01 01 14 7 B 14 M 15 01 0 S . , 11 01 01 S 26 M Q 0 13 M M Q 23 M Q 14 10 01 01- 01 16 16, 4 B M S 5 14 0 S 20 B 14 Q 5 M Q 0 23 B 14 M 5 01 71 12 M M S 12 T 01 S 14 01 14 0 7 T M Q 17 01 M M 7 T M M 20 14 0 5 18 14 0 0 23 4 4 01 M S 23 M 0 M 28 01 M Q 29 T M S 8 M M N 19 T M 0 7 M 0 S 28 T M M 6 M 0 .0 36 M M 0 32 17 1 14 M 14 14 5 4 M 01 S 68 18 4 M 14 S 19 M N 60 0 01 M Q 16 M M 14 21 N M M 16 N S 0 93 N 0 S 23 M 106 Q 0 M 0 0 20 N 0 M 71 14 0 01 19

6 1 B 14 M 14 19 1 N M M 8 M M II 15 'S M M 10 20 4 N N S 18 a N 14 0 19 6A 1 M S M 9 14 M N 7 N Q S 19 7 1 M M M 16 M Q 0 19 14 0 M 17 8 4 M S M Q 9 * Code, Along Length N M M 6 B Bottom 8A 4 01 M S 19 14 Midlength M 01 0 14 T Top 01 M M 14 Code, Along Width N Weld centerline 9 4 5 M 9 13 Q Halfway between weld centerline and bond line B 01 Q 7 :ode In Thickness 5 N 01 8 S Surface 01 N S 7 0 Quarterpoint N M Q 16 M Nidthickness N 01 M 7 01 Q S 15 Conversion Factors N 0 Q in M 0 M 20 1inch=25.4,,,n N S 9 1 ft-lb 1.356 J 'S M 0 8 'S M 01 5

10 4 M 14 9 11 M 01 0 7 N N M 9 N 0 S 12 M Q 0 .11 N 0 M 16

cooling rate, etc., did not correlate with the incidence of the fatigue tests and the low-temperature impulse tests are these separations. discussed separately in the following. 10. Tension tests of material containin'g separations At about 1.3 million cycles, a large crack was noticed indicated no loss of yield and tensile strength; however, several inches from the joint in detail 36BSR. This crack, elongation and reduction-of-area values were reduced. which initiated at a nick caused by knocking off a tack- welded strongback, was repaired and did not affect the test PHASE II results. Consequently, it is not mentioned in the discussion of the fatigue-test results. It is mentioned here only to The behavior of the two specimens is summarized by emphasize the need for properly repairing any surface the chronology of events given in Table 7. The results of imperfections caused by removal of the strongbacks. 34

gory B is plotted in Figure 27. 'This category applies to Fatigue-Test Results all structural steels and covers several different types of Stress/Lite Relationships details including straight and tapered joints with the weld reinforcement removed. The allowable stress-range line is The fatigue-test results for the four joint details are 95 percent confidence limit summarized in Table 23. For joints that were repaired, the intended to represent the lower (12). (More precisely, this allow- number of cycles applied both before and after repair is for the covered details able stress-range line is a straight line drawn two standard listed. The results are also plotted on an SN diagram in deviations below the mean line for the data.) The tests Figure 27. Stress range is used as the stress parameter in that were conducted to establish the line included 81 tests this diagram because past studies (12, 13) have shown of conventionally welded joints with a taper in width (12). that it is the major parameter affecting fatigue life. Type The stress-concentration factor for these joints was reported of steel and minimum stress have a much smaller effect No tests were conducted on joints with a Therefore, data for both specimens were in- to be 1.1 (12). (12, 13). taper in thickness or on any type of joint welded by the cluded on a single plot. When 50 percent of the flange was cracked, the detail was considered to have failed. For electroslag process (12). details 588C and 36B5R, the estimated numbers of cycles The AASHTO allowable stress-range line for Category C necessary to propagate the crack to 50 percent of the is also shown in Figure 27 for comparison with the results flange area were added to the listed lives before they were for detail 588C. The tests conducted to establish this line plotted in Figure 27. These estimated additional cycles did not include any straight or tapered joints with the weld Such joints were assigned to are 20,000 and 10,000, respectively. reinforcement in place (12). The AASHTO allowable stress-range line (6) for Cate- a lower detail category than similar joints with the rein-

TEST TEMPERATURE. C TEST TEMPERATURE, C - -20 -10 0 10 20 -0 -10 0 10 20

-32 0

BOTTOM 80 0 MIDDLE

TOP . 100 00

70 70 90 -9 0 0 80 8 & S 0 - a a 8 -80

- 50 co

- 3 0 0 0 30

0 0 0 Ell 000 00S 20 00

0 - 0 $ 0 - 10 0 00 000

II 3 -20 0 20 40 60 80 - -20 0 70 TEST TEMPERATURE. F TEST TEMPERATURE. F Figure 20. Individual Charpy V-notch energy absorption values for Figure 21. Individual Charpy V-notch energy absorption values for HAZ in weldment No. 1. HAZ in weidment No. 3.

35

TABLE 20 TABLE 21 SUMMARY OF HEAT-AFFECTED-ZONE CVN ENERGY RESULTS OF DROP-WEIGHT TESTS OF ELECTROSLAG ABSORBED (AWS METHOD) WELDMENTS

CVN Energy Plate Nil-Ductility Temperature, 'F Weldmcnt Thickness, Along in Absorbed, ft-lb welciment Steel Thickness, Specimen Heat-Affected Weld No. in. Length' Thickness' O'9' 40'F' - No. Grade inches Orientation Base Metal Zone Metal 1 1 N N 7 16 1 A36 1 N 30 30 0 2 1 M N 11 20 3 A36 4 P - 20 -40 N -20 20 -40 3 4 B 0 26 - N 0 21 37 6 A588 1 P 40 -10 0 T 0 17 - 8 A588 4 - P - 0 0 4 4 M 24 36 Q N 60 0 10

4 K Q 25 30 * P: Plane of specimen parallel to plate surface. N: Plane of specimen normal to plate surface. 6 Same welding conditions as weldnent 19 Conversion Factors

7 1 N M 10 43 1 inch = 25.4 mm - C = 5/9 ('F - 32) 8 4 M Q 8 15

8A 4 N Q 12 -

9 4 N Q 11 17

10 4 M Q 9 10

11 1' side N N 31 32 2" side M 0 27 20

12 2' side M Q 21 30 4' side N Q 11 18

13 1 M N 7 17 / 14 1 M N 17 23

15 4 - B 0 8 - N Q 9 12 T Q 9 -

16 4 B Q 12 - M Q 12 23 TEST TEMPERATURE, C T Q 8 0 50 tOO 800 17 1 N N 8 25 1 II I I I I II I I I I

18 4 M Q 17 18 £ A A36 BASE METAL I 19 1 N N 16 23 700 OHAZ A 0 20 4 M Q 14 17 0 WELD METAL A * Code: Along Length + Average of 5 specimens B Bottom ++ Average of 3 specimens 6 M Midlength T Top Conversion Factors Code: In Thickness 0 Quarterpoint 1 inch = 25.4 mm M Midthjckness 1 ft-lb 1.356 J AS 'C = 5/ 9 ('F - 32)

'C 2 0 Maz U' TABLE 22 RELATIONSHIP BETWEEN CHARPY V-NOTCH AND 2 DROP-WEIGHT NDT RESULTS

Plate EnergyAbsorbed (ft-lb) at NDT - Weldment Steel Thickness, Base neat-Affected Weld No. Grade in. Metal Zone Metal

1 A36 1 30 14 21 1

3 A36 4 25 23 6

6 A588 1 28 14 15 -50 300 8 A588 I 33 9 9 TEST TEMPERATURE. 'F Conversion Factors

1 inch = 25.4 mm 0 1 ft-lb = 1.356 J Figure 22. Dynamic tear-test results for weidment No. 18.

36

TEST TEMPERATURE, C TEST TEMPERATURE, C U DO

SHEAR ENERGY A A588 BASE METAL 800 800 — a a WELD METAL

1 ft-lb = 1.356 700 700

600 600 .0 .0

0 500 500 0 cc

!0 40O 20 Irz

300

4

200 cr 200 4

(4

100 G 100 80

0 00 0 100 -50 0 50 100 150 200 250 TEST TEMPERATURE, F TEST TEMPERATURE, Figure 23. Dynamic tear-test results for weidment No. 9. Figure 24. Dynamic tear-test results for weidment No. 15.

TEST TEMPERATURE, C 0 50 100 I I I I I

SHEAR ENERGY A A588 BASE METAL 800 £ OHAZ a a WELD METAL

700 1 ft-lb= 1.356J

0 0-0 .0600 0

o 0 -0 .-.,,.,.

Ct 300

200

A 100 'a

I I

0 50 100 1 150 200 250 300 TEST TEMPERATURE, F Figure 25. Dynamic tear-test results for weidment No. 16. 37 forcement removed to account for the effect of the stress concentration caused by the reinforcement. Two out of five B details (detail 588CR is considered a B detail because the weld reinforcement was ground off as part of the repairs) failed at lives well below the life of 2 million cycles corresponding to the AASHTO allowable stress-range line. Specifically, detail 588B had a life of 1,243,000 cycles and detail 36BS had a life of 621,000 cycles. Since this AASHTO line represents a lower 95 percent confidence limit, the probability that these two points would be so low is very small. The results for the detail 588C also fell well below the AASHTO line for Category C and would have a very low probability of occurrence. This indicates that the AASHTO Categories B and C do not adequately represent the data from the present tests on electroslag-welded tapered joints. There are five possible explanations for these low results: (1) because of poor weld quality or unsatisfactory test methods, the results from the present tests are not typical of results for similar tapered electroslag welded joints; (2) ::U electroslag welding results in a lower fatigue strength than A. thd shielded metal-arc or submerged-arc welding when used in any joint configuration; (3) joints with a thickness taper have a lower fatigue strength than straight joints made by the same welding process; (4) electroslag welds have a greater fatigue notch sensitivity than shielded-metal-arc or submerged-arc welds and, therefore, have a lower fatigue strength in joints involving significant stress concentrations; and (5) because of a size effect, the present results for large joints are lower than the results for similar smaller joints on which the AASHTO specifications are based. Each of these possibilities is discussed in the following.

Reinforcement

5 -- - 4 '

3 0 -. B. Etched in nital.

Figure 26. Grain-boundary separation as observed in weidment No. 3 (X 100).

TABLE 23 FATIGUE TEST RESULTS Before Repair After Repair Loading, ksi Cracked Cracked Detail Stress Range Max Stress Cycles Area, % Cycles Area,

588B 18 27 1.243,000 50 - - 588c 13 23 618,000 25 625,000 0

36BN 18 20 2,528,000 0 - - 368S 18 20 621,000 55 1,907,000 30 As part of the repair, the weld reinforcement was ground of f to CYCLES TO FAILURE a smooth Contour. Therefore, the repaired details are Classified Figure 27. Fatigue data on electroslag welded joints with a taper as a Category B detail. in thickness. 1 ksi = 6.895 MPa 38

The first possibility is that the results from the present with and without the weld reinforcement in place are tests are not typical of results for similar tapered electroslag identified separately in the figure. The two data points welds because of poor weld quality or unsatisfactory test marked "X" designate the special results discussed in the methods. The present results are the only known data on Phase II report (3). tapered electroslag-welded joints. Therefore, it is not pos- All the regular data shown in Figure 28 are above the sible to say conclusively whether the present results are appropriate AASHTO allowable stress-range lines except typical of such joints. However, investigations of the welds for three data points at the high end of the range, which indicated that they were acceptable according to the are slightly below the appropriate lines. This suggests that requirements of the AWS specifications (5) and the addi- the straight electroslag-welded joints do not have signifi- tional acceptance criteria for electroslag welds recom- cantly lower fatigue lives than similar joints made by other mended in Phase I. Similarly, the test methods are con- welding processes. Furthermore, the researchers who per- sidered satisfactory; there were three independent checks formed these tests (14, 15) stated the same conclusion. on the magnitude of the loading. Consequently, the present The third possibility is that joints with a thickness taper fatigue results are considered representative of those of have a lower fatigue strength than straight joints made similar tapered electroslag-welded joints. by the same welding process. The measured stress concen- The second possibility is that electroslag welding results tration factor for the tapered -joint was about 2. This is in a lower fatigue strength than shielded-metal-arc or close to the theoretical value of 2.15 calculated (16) by a submerged-arc welding when used in any joint configura- finite-element analysis for a joint with a 1 in 21/2 taper and tion. To evaluate this possibility, the results of all known a ratio of plate thicknesses of 11/2 . Furthermore, the fatigue tests (14, 15) on full-thickness electroslag-welded decay rate of stress concentration is considerably less than joints without a taper are plotted in Figure 28. The results the rate for other details, such as a cover-plate end or include five tests recently conducted (this work is not part stiffner (16). Both the magnitude of the stress-concentra- of NCHRP Project 10-10) at the U.S. Steel Research tion factor and the decay rate affect fatigue life. Laboratory; these results have not previously been pub- Theoretical stress-concentration factors reported in the lished. The same welding procedures were used in these literature for certain other details are listed, as follows, five specimens as were used for specimen 36. Data for to permit a comparison with the factor for the tapered specimens with plate thicknesses ranging from 11/4 to 2 in. joint: (32 to 51 mm) are included in Figure 28. Data for joints Straight butt joint with reinforcement off:1.0. Straight butt joint (17) with reinforcement on (60 degree angle at the toe of the reinforcement) :1.3 to Reinforcement Special Off On Points 1.8. Ref14 0 Joint with a taper (12) in width (1 in2½ slope):1.1. 60 Ref 5 0 X Unpublished £ A Transverse stiffener (16, 18, 19) :2.2 to 4. —Test Discontinued 50 .. The theoretical factors for a particular type of detail vary considerably with various geometric parameters such as the 40 - U 0 slope of the weld or weld reinforcement. Furthermore, the actual stress concentrations probably vdry even more as S .— 0 0 o 0 a result of unintentional variations in weld contour and 30 • 0 0 other geometric parameters. Nevertheless, the theoretical factors in combination with the decay rates provide an - 0— CategoryB ._ indication of the relative severity of different details. 2 The stress-concentration factor of 2 and the low decay 20 rate for the joint with a taper in thickness indicate that this (I) CategoryC V) detail is considerably more severe than a straight joint or w a joint with a taper in width. In fact, the severity of the I— x. U) thickness taper approaches that of a stiffener, which has X A— a higher stress-concentration factor and also has a higher decay rate. Therefore, the joint with a thickness taper would be expected to have a lower fatigue strength than a tO straight joint or a joint with a width taper. 9 A few fatigue results are available (16, 20, 21) on 8 conventionally welded joints that are tapered in thickness. 7 These results are shown in Figure 29. Six results are for girders with a 1 in 21/2 thickness taper in the flanges (16). 6789100 2 3 4 5 678910 2 3 4 5 The ratio of the thicknesses of the joined flange plates was CYCLES.TO FAILURE 1.5 compared with a more severe ratio of 2.0 for the

Figure 28. Fatigue data for electroslag welded joints with no girders in the present study. All but two of the data in taper. Figure 29 are above the appropriate AASHTO allowable- 39 stress-range lines. Thus, these limited data suggest, but do 60 not prove, that a conventionally welded joint with a taper in thickness and the weld reinforcement removed can be 50 safely classifie1 as a Category B detail, even though it may have a fatigue strength less than that of a straight joint. 40 The fourth possibility is that electroslag welds have a greater fatigue notch sensitivity than shielded-metal-arc or submerged-arc welds and, therefore, have a lower fatigue 30 U, strength in joints involving significant stress concentrations. S Quantitative values of the fatigue notch sensitivity for electroslag welds are not available. However, a recent study of notched electroslag-welded specimens gives an indirect indication of notch sensitivity (22). This study suggests that the electroslag welds have a higher notch sensitivity than submerged-arc welds if the notch is at a critical location in the weld. Thus, the stress concentration caused by the thickness taper may have a somewhat larger effect on an electroslag weld than on a submerged-arc weld. 10 The fifth possibility is that, because of a size effect, the present results are lower than the results for similar smaller joints on which the AASHTO specifications are based. Because weld quality tends to decrease (22) and the possibility of flaws tends to increase with weld size, it is 6 likely that the present results are slightly lower than the 6 78910 2 3 4 5 6789106 results for smaller tapered joints. But it is unlikely that CYCLES TO FAILURE these effects alone are sufficient to account for the low Figure 29. Fatigue data for conventional welded joints with a fatigue lives that were observed. taper in thickness.

Crack-Growth Relationships The growth of surface cracks in the specimens was I I 1 1111 1 observed during the tests and is recorded in the Phase II report. The crack-growth rate for detail 588B was deter- do H = 1.68 1010 :. mined from these observations and from a study of the _jN- (AK)33 crack surface after failure. The stress-concentration factor and decay rate were accounted for in calculating the growth rate. Sufficient data were not available to permit a similar determination for the other details. The crack- I•4j growth-rate data for detail 588B includes values for crack depths ranging from 0.25 to 1.05 in. (6.4 to 27 mm). The corresponding tK values are 15.1 and 25.8 ksi\TIii'. (16.6 d. and 28.4 MPaV). Values of da/dN corresponding to 1.09 10-10 (AK)33 these two values of zK are plotted in Figure 30 for com- parison with crack-growth rates obtained from small. X-GIRDER DATA wedge-opening loading (WOL) specimens of structural steels. As discussed earlier, similar WOL data for the LOAD STRESS weld metal, the bond line, and the HAZ of electroslag STEEL RANGE, lb RATIO (R) REF welds agreed with these base-metal data except, in some A36 2500 0.07 4 cases, where lower growth rates were obtained at the lower 3000 0.06 4 10 3320 0.11 5 end of the curves for the weld specimens. At the larger -I.' A588-A 2590 0.07 4 crack depths, the growth rate from the girder test agreed -I 3105 0.06 4 well with those previously obtained for base metal. At .Ig 1T-WOL SPECIMEN the smaller crack depths, the growth rates from the girder AIR ENVIRONMENT test were somewhat higher than those previously obtained FREQUENCY: 300 cpm for the base metal. This suggests that the actual stress concentration factor at the small depth may have been 1 inch = 25.4 mm 1 lb = 4.448N larger than the factor used in the analysis. 1 ksi./inch = 1.1 MPa v'rn I 1111111 I 111111 Low-Temperature Impulse-Test Results 10 20 30 40 60 80 100 Table 24 summarizes the low-temperature impulse-test STRESS-INTENSITY-FACTOR RANGE, AK, ksi y'iflch results for each detail. The specimens contained large Figure 30. Crack-growth rate for detail 588B. TABLE 24 C IMPULSE TEST RESULTS Max* Ave. Mm/Max Stress Net- Approx. Crack Stress Impulse Max Stress Rate, Section Impulse Cracked Depth, Crack Coefficient, Temperature Stresses, Failure Intensity, K -ksi/sec Stress, Detail No. Area, % a, in. pe C Reference OF c, ksi Stress, ksi ksiV'in. Max Ave ksi 588B 1 50 6.0 3 2.9 13 0 9/26 NF 327 10.7 6.8 52 2 50 6.0 3 2.9 13 -36 9/22 22 278 10.2 6.5 44 588CR 1 0 0 - - - -2 9/26 NF - 10.6 6.7 26 2 0 0 - - - -39 9/22 NF - 10.2 6.4 22 36BN 1 0 0 - - - -31 7/20 NF - 18.3 11.6. 20 2 0 0 - - - -28 7/34 NF - 18.1 11.5 34 3 20 0.42 4 1.3 9 -27 7/20 NF 30 16.8 10.7 25 4 20 0.42 4 1.3 9 -32 7/34 NF 51 18.6 11.9 43 5 20 0.42 4 1.3 9 -60 6/34 NF 51 18.1 11.5 43 6 20 0.42 4 1.3 9 -59 7/35 NF 52 18.2 11.6 44 7 40 0.75 4 1.5 9 -31 6/18 NF 41 17.2 11.0 30 8 40 0.75 4 1.5 9 -33 6/26 NF 60 16.7 11.2 43 36B5R 1 0.2 0.2 1 0.72 9, 13 -26 7/19 NF 11 18.0 11.5 19 2 0.2 0.2 1 0.72 9, 13 -23 7/35 NF 20 18.7 11.9 35 3 5 1.3 2 0.85 9, 13 -27 7/19 NF 33 16.9 10.5 20 4 5 1.3 2 0.85 9, 13 -30 7/35 NF 60 18.8 12.0 37 5 5 1.3 2 0.85 9, 13 -63 6/34 NF 58 18.3 11.6 36 6 5 1.3 2 0.85 9, 13 -61 6/34 NF 58 18.3 11.7 36 7 30 3.8 3 1.8 13 -34 . 6/20 NF 124 18.6 11.8 29 8 30 3.8 3 1.8 13 -33 6/28 28 174 16.7 12.1 40 1 ksi = 6.895 MPa * Maximum impulse stress divided by the proportion of the tension-flange O C 5/9 = (°F - 32) area that remains unccked, un. = 25.4mn - 1 ksiVT= 1.0998 MPaV'm __ TyP a) a Type 1 a Type 2 Pc 3 4 Liv - r'' " i 41 fatigue cracks at the welded joints when they were sub- It is significant that the net-section stresses in most of the jected to the low temperature impulse loadings. The stress tests were well above the yield strength of the material at values listed in Table 24 are nominal bending stresses at the the test temperature. This indicates that gross yielding toe of the weld and exclude the effects of stress concentra- occurred in the presence of a sharp crack and that the tion due to joint geometry and any cracks present at this fracture event was load-limit rather than fracture-toughness location. (As used here, weld toe refers to the intersection dominated. between the sloping portion of the weld and the horizontal The estimated maximum stress intensity for each impulse face of the 2-in.-thick flange plate.) is listed in Table 24. This stress intensity corresponds With about 50 percent of the flange area cracked, the to the maximum nominal stress for that impulse and was A588-steel electroslag-welded tapered joint did not fail calculated by standard fracture mechanics formulas iden- when subjected to a maximum impulse stress of 26 ksi tified in the table. The estimated stress intensities, K, in (179 MPa) at OF (-18 C). This is close to the AASHTO details 588B and 36BSR at the instant of failure were 278 maximum allowable stress of 27 ksi (186 MPa) for this and 174 ksiV. (306 and 191 MPa\/rn), respectively. steel. At —30 F (-34 C), this severely cracked detail These results are for a temperature of about —30 F was still able to withstand a stress of 22 ksi (152 MPa)- (-34 C). The values greatly exceed the limits for valid 80 percent of the maximum allowable stress—before K1 values at —30 F for electroslag welds and A36-steel failure. plates shown in Figure D-4 of the Phase I report. This With about 20 percent of the flange area cracked, the range is 47 to 60 ksi\/in. (52 to 66 MPa\/ii). The A36-steel electroslag-welded tapered joint did not fail when maximum stress intensity occurring in detail 36BN when subjected to a maximum impulse stress of 35 ksi (241 failure occurred in detail 36BSR was at the top of this MPa) at —60 F (-51 C) This is close to the specified range of values. is the critical (failure) stress intensity minimum yield strength of 36 ksi (248 MPa). With about K1 under plane-strain conditions and represents the minimum 40 percent of the flange area cracked, the joint did not fail values from when subjected to a maximum impulse stress of 26 ksi fracture toughness of a material. The K1 (179 MPa) at —30 F. With about 30 percent of the flange Phase I were obtained with fatigue-precracked bend area cracked, the repaired A36-steel electroslag-welded specimens. tapered joint was able to withstand a maximum impulse The large K values estimated for the impulse tests reflect stress of 28 ksi (193 MPa) at —30 F before failure. This the ductile toughness behavior of cracked members, which stress is well above the AASHTO allowable value of 20 ksi have net-section stresses exceeding the yield strength. (138 MPa) and is about 75 percent of the specified mini- The failure load for such members is governed by limit-load mum yield strength. conditions rather than linear-elastic fracture mechanics.

CHAPTER THREE

INTERPRETATION AND APPRAISAL OF FINDINGS

BACKGROUND protracted thermal cycle, result in an extremely large HAZ and a large unrefined grain size in the weld metal. The purpose of NCHRP Project 10-10 was to develop This has caused concern about the use of electroslag weld- acceptance criteria for the use of electroslag weldments in ing under the conditions of fatigue and moderate dynamic bridges. Toward this purpose the available literature was loading encountered in bridges. reviewed and an extensive welding and testing program this process has had a very successful history of was conducted. The primary emphasis was placed on application in buildings and process equipment subject to determining (1) whether electroslag weldments are suitable moderate dynamic loading such as blast furnaces, basic- for main structural components in bridge applications and oxygen furnaces, slag pots, etc. The two greatest areas of (2) what procedure and control testing should be required concern for transportation officials considering this process if electroslag welding is used in bridgçs. for bridge applications have been the serviceability of Electroslag welding is a relatively new welding process electroslag weldments in situations of fatigue loading and with inherent economic advantages. As used in structural of moderate impact loading at low temperatures. Project welding, it also has other advantaes over conventional 10-10 was designed to evaluate these concerns. submerged-arc, shielded-metal-arc, and flux-cored-arc weld- ing in that, generally, the weld metal contains fewer entrapped inclusions, distortion is minimized, and a favor- ELECTROSLAG WELDMENTS able residual-stress pattern is developed. Metallurgical factors, particularly the effect of the high heat input and With respect to the suitability of electroslag weldments 42 in A36 and A588 steels for main structural components, AASHTO maximum allowable bending stresses even when the following comments are applicable: 20 percent or more of the tension-flange area is cracked. In Adequate strength and ductility across properly made service, a bridge member with a crack this size would soon electroslag weldments can be readily obtained. In fact, fail by fatigue regardless of the level of fracture toughness the properties of as-welded electroslag weldments, par- in the joint. Thus, an increase in the fracture toughness of ticularly yield strength and hardness, more closely match the joint would not significantly affect the service per- those of the base metal (for structural steels with minimum formance of such members. yield points up to and including 50 ksi or 345 MPa) than A large crack developed in one of the girder speci- do the properties of weldments produced by arc welding by mens at a nick caused by knocking off a tack-welded conventional processes. strongback. This crack did not affect the test results, but is The fatigue crack propagation behavior of electro- mentioned here to emphasize the need for properly re- slag-welded laboratory specimens in Phase I was satis- pairing any surface imperfections caused by removal of factory. These fatigue-crack growth-rate studies confirmed the strongbacks. previous studies; that is, in properly made electroslag One area of concern is the occasional occurrence of weldments, the fatigue properties of the weld metal and grain-boundary separations (microcracks) that are below the associated HAZ are as good as, or superior to, those of the level of detection when inspected in accordance with the structural steels used in bridges. In contrast, the fatigue the provisions of the current AWS Structural Welding lives of the electroslag-welded tapered joints tested in Code. In procedure qualification testing, the four side-bend Phase II were lower than expected from previous studies tests specified by AWS Dl. 1 would establish the presence and did not meet the AASHTO fatigue requirements for or absence of grain-boundary separations and limit the the applicable detail categories—Category B for joints with maximum acceptable size to ½ in. (3.2 mm). The occur- the weld reinforcement removed and Category C for rence of separations cannot be fully explained, although joints with the weld reinforcement in place. The reasons hydrogen-induced propagation from micropore initiation for this could not be definitely established. However, it is sites is a possible mechanism. However, to put things in likely that electroslag welded joints with a taper in thickness proper perspective, all weldments are subject to imperfec- have a significantly lower fatigue strength than straight tions, which may or may not be detected under production joints and should not be included in Category B. On the testing conditions. Further research is necessary to estab- other hand, there is little evidence to indicate that straight lish the effect of these grainboundary separations on the electroslag-welded joints should not be included in Cate- fatigue life of electroslag weldments. gory B or that they have a lower fatigue strength than straight joints made by other welding processes. Thus, ACCEPTANCE CRITERIA more fatigue tests of tapered and straight electroslag- With respect to what procedure and control testing welded joints are needed to adequately define the fatigue should be required, the AWS Dl.1 Structural Welding Code strength of such joints. However, the present study sug- was accepted in this study as representing essentially the gests that the following detail categories would be appro- minimum requirements. In addition, supplemental testing priate electroslag-welded joints, with the weld reinforce- requirements, which may be needed to assure adequate ment removed: Category B for straight joints and Category performance in bridge applications, were considered. In C for tapered joints. The present study also suggests that other words, reduction or elimination of current require- electroslag-welded joints with the reinforcement in place ments for strength, ductility, and soundness was not con- should not be used in fatigue applications. sidered. In this context the following comments summarize Adequate notch toughness for AASHTO Zones 1 and the recommendations proposed at the conclusion of Phase I 2 service can be obtained by using commercially available of Project 10-10: materials and practical production-welding procedures. Judging from the CVN tests and other notch-toughness The minimum requirements of the AWS Dl. 1 Struc- measurements discussed in Chapter Two, adequate prop- tural Welding Code, including Appendix C, should be met. erties at service temperatures as low as —30 F (-34 C) For tension members in bridge applications, notch- can be obtained with commercially available materials and toughness requirements of both the weld metal and the procedures. For AASHTO Zone 3 service (lowest service HAZ should be made mandatory. Under the current AWS temperature —60 F) under the present state of the art, Dl. 1 Code, impact requirements of electroslag weldments electroslag weldments in the as-welded condition do not are corsidered supplementary—minimum average CVN appear suitable for dynamic applications, such as bridges, energy absorption of 15 ft-lb at OF (20 J at —18 C), with because of low notch toughness in the weld heat-affected no values under 10 ft-lb (14 J), based on testing five zones. specimens and discarding the highest and lowest values. Limited fracture-mechanics-type (K1 ) tests of 1- Because the HAZ is large and extends essentially directly and 4-in.-thick electroslag weldments of A36 and A588 across the plate thickness (electroslag weldments are gen- Grade A steels suggest that electroslag weldments can be erally made with a square-groove butt joint preparation), made having fracture-toughness properties that are ade- it is recommended that impact requirements be specified in quate for use in bridges. This conclusion is supported by this area. On the basis of the results of the present study, the Phase II results, which showed that full-scale girder CVN impact tests appear to be the most practical for specimens can withstand stresses at —30 F, well above the production and control testing. It is also recommended 43 that five specimens be obtained from the quarter-thickness Changes in process variables cause changes in micro- location and notched in the grain-coarsened region of the structure and macrostructure that, in turn, cause changes HAZ near the bond (overlength blanks should be obtained in properties. Assessment of the degree of variation in and etched to determine the location of the notch). These items such as voltage, amperage, etc., which require specimens should be tested and averaged by the AWS requalification, is covered in AWS Dl. 1-76. Their adequacy method (discarding the highest and lowest values), and was not assessed in Project 10-10. the average value should meet the minimum current Nondestructive inspection techniques, particularly for AASHTO requirements for the steel being welded. For detection of grain-boundary separations, were also not example, for A36 steel for Zone 2 service, the HAZ would assessed in Project 10-10. have to meet 15 ft-lb at 40 F (5 C). Observations in this study indicate that there is extreme RECOMMENDATIONS AND FUTURE WORK scatter in the CVN test results for electroslag welds; how- Because of the uncertainties regarding the fatigue ever, averages are fairly consistent. With respect to test strength of electroslag welds raised by the present program, location in the weld metal, the center appears to exhibit the additional research is needed. Specifically, fatigue tests lowest CVN values, with the properties improving in all should be performed on representative full-thickness directions from the weld center. Therefore, it is recom- tapered and straight electroslag-welded joints to determine mended that CVN test specimens be obtained from the whether present results are typical of similar electroslag- quarter-thickness location on the weld centerline away welded joints and to establish the appropriate AASHTO from the weld start and stop. This area is fairly representa- detail category for such joints. Similar tapered shielded- tive of the over-all weld and is in a region where major metal-arc and submerged-arc-welded joints should also be cracks could propagate. From a practical standpoint, this tested to show whether there is a significant difference in is a good point for locating specimens. The quarter point the fatigue strength of tapered joints made by the various at the weld centerline is easily located and is, the standard welding processes. A few girder or large beam specimens location for testing other types of welds. Also the quarter with tapered and straight electroslag-welded joints should point is used for testing plates where properties also vary be tested to show that the results for the full-thickness from surface to center. Testing of the extreme ends of specimens are applicable to such specimens. welds is not considered practical. For example, to test There is also a need for further investigations of the the weld center at the weld top (presumably the worst metallurgical characteristics of electroslag welds that might location), eight welds would have to be made to obtain cause low fatigue strengths, such as grain-boundary separa- sufficient specimens for one qualification test. The weld- tions that result in microcracks. Restraints may contribute centerline, quarter-point location is recommended for to this microcracking and therefore cause more problems testing; establishment of the fracture toughness minimum in large, fabricated bridge members than in small laboratory is also recommended at that location, taking into account specimens. Although it does not appear that grain- that the values that would be obtained at other locations boundary separations contributed to the low fatigue results may be less. from the present tests, they could cause reductions in Because of the observed scatter of individual test results, fatigue strengths. If it is possible to produce separations in it is suggested that eight weld-metal specimens be tested, full-thickness specimens like those discussed in the preced- rather than the normal five, and that the highest and lowest ing paragraph, it would be desirable to include such values be discarded. The current AWS P1.1 supplemental specimens in the fatigue program discussed earlier. This toughness requirements of 15 ft-lb average at 0 F should be would provide a direct measure of the effect of the separa- adequate for AASHTO Zoncs 1 and 2 service. The more tions on the fatigue strength of tapered and straight joints. stringent requirements for the weld metal are predicated Nondestructive inspection methods applicable to electro- on the greater likelihood of encountering imperfections slag weldments comprise another area where additional below detectable levels in weld metals. research is needed.

RE FERENCES

BENTER, W. P., JR., "Electroslag Weldments in SCHILLING, C. G., and KLIppsTEIN, K. H., "Fatigue Bridges." NCHRP Research Results Digest 74 (June and Low-Temperature Impulse Tests of Electroslag 1975) 23 pp. Welded Girders, Phase 2, Final Report." NCHRP Project 10-10 (June 30, 1978). BENTER, W. P., JR., KONKOL, P. J., KAPADIA, B. M., "Electroslag and Electrogas Welding," Chap. 7 Vol- SHOEMAKER, A. K., and SOVAK, J. F., "Acceptance ume Two, Seventh Edition, Welding Handbook, Criteria for Electroslag Weldments in Bridges— American Welding Society (1978). Phase 1, Final Report." NCHRP Project 10-10 (Apr. AWS D1.1 Rev. 2-77, "Structural Welding Code." 1, 1977). American Welding Society (1977). 44

6. American Association of State Highway and Trans- 15. HARRISON, J. D., "Fatigue Tests of Electroslag Welded portation Officials, Standard Specifications for High- Joints." Metal Construction and British Welding way Bridges (1977). Journal Vol. 1, No. 8 (Aug. 1969). 7. CUDNEY, G. R., "Stress Histories of Highway Bridges." 16. BOYERS, K. D., ET AL., "Determination of Tolerable J. of Structural Div., ASCE, Vol. 94, No. ST12 (Dec. Flaw Sizes in Full Size Welded Bridge Details." Fritz 19.68). Engineering Laboratory, Report No. 399-3(76), 8. American Association of State Highway and Trans- Lehigh University (Dec. 1976). portation Officials, Standard Specifications for Welding 17. LAWRENCE, F. V., "Estimation of Fatigue-Crack of Structural Steel Highway Bridges (1978). Propagation Life of Butt Welds." Welding Research 9. BLAKE, G. T., "Structural Testing Systems." U.S. Supplement (May 1973). Steel Corporation, Research Laboratory Report 18. FISHER, J. W., ET AL., "Fatigue Strength of Steel 57.019-910(1) (June 11, 1971). Beams With Welded Stiffeners and Attachments." 10. SCHILLING, C. G., ET AL., "Low Temperature Tests of NCHRP Report 147 (1974) 85 pp. Simulated Bridge Members." U.S. Steel Corporation, 19. ZETTLEMOYER, N. and FISHER, J. W., "Stress Gradient Research Laboratory Report 97.021-001 (3) (Dec. 31,, Correction Factors for Stress Intensity of Welded 1972). Stiffeners and Cover Plates." Welding Research 11. SCHILLING, C. G., ET AL., "Low Temperature Tests of Supplement (Dec. 1977). Simulated Bridge Members." I. of Structural Div., 20. MUNSE, W. H., Fatigue of Welded Steel Structures. ASCE, Vol. 101, No. ST1 (Jan. 1975). Welding Research Council, New York City (1964). 12. FISHER, J. W., ET AL., "Effect of Weldments on the 21. STALLMEYER, J. E., and MUNSE, W. H., "Behavior of Fatigue Strength of Steel Beams." NCHRP Report Welded Built-Up Beams Under Repeated Flexural 102 (1970) 114 pp. Loads." Dept. of Civil Engineering Report, University 13. SCHILLING, C. G., and KLIPPSTEIN, K. H., "Fatigue of of Illinois (Sept. 1966). Steel Beams by Simulated Bridge Traffic." J. of Struc- 22. CULP, J. D., "Fracture Toughness and Fatigue Prop- tural Div., ASCE, Vol. 103, No. ST8 (Aug. 1977). erties of Steel Plate Butt Joints Welded by Submerged 14. NOEL, J. S., and Topnc, A. A., "Static, Fatigue, and Arc and Electroslag Welding Procedures." Research Impact Strength of Electroslag Weidments." Center Report No. R-1011, Testing and Research Division, for Highway Research, University of Texas, Research Michigan Department of State Highways and Trans- Report 157-iF, Project 3-5-71-157 (Dec. 1972). portation,(May 1976). THE TRANSPORTATION RESEARCH BOARD is an agency of the National Research Council, which serves the National Academy of Sciences and the National Academy of Engineering. The Board's purpose is to stimulate research concerning the nature and performance of transportation systems, to disseminate information that the research produces, and to encourage the application of appropriate research findings. The Board's program is carried out by more than 150 committees and task forces composed of more than 1,800 administrators, engineers, social scientists, and educators who serve without compensation. The program is supported by state transportation and highway departments, the U.S. Department of Transportation, and other organizations interested in the development of transportation. The Transportation Research Board operates within the Commission on Sociotech- nical Systems of the National Research Council. The Council was organized in 1916 at the request of President Woodrow Wilson as an agency of the National Academy of Sciences to enable the broad community of scientists and engineers to associate their efforts with those of the Academy membership. Members of the Council are appointed by the president of the Academy and are drawn from academic, industrial, and govern- mental organizations throughout the United States. The National Academy of Sciences was established by a congressional act of incorpo- ration signed by President Abraham Lincoln on March 3, 1863, to further science and its use for the general welfare by bringing together the most qualified individuals to deal with scientific and technological problems of broad significance. It is a private, honorary organization of more than 1,000 scientists elected on the basis of outstanding contribu- tions to knowledge and is supported by private and public funds. Under the terms of its congressional charter, the Academy is called upon to act as an official—yet indepen- dent—advisor to the federal government in any matter of science and technology, although it is not a government agency and its activities are not limited to those on behalf of the government. To share in the tasks of furthering science and engineering and of advising the federal government, the National Academy of Engineering was established on December 5, 1964, under the authority of the act of incorporation of the National Academy of Sciences. Its advisory activities are closely coordinated with those of the National Academy of Sciences, but it is independent and autonomous in its organization and election of members. TRANSPORTATION RESEARCH BOARD NON.PROFIT ORG. National Research Council U.S. POSTAGE 2101 Constitution Avenue, N.W. PAID Washington, D.C. 20418 WASHINGTON, D.C. ADDRESS CORRECTION REQUESTED PERMIT NO. 42970

0 V N >- = 0 LL. i 0 >

a. U.' 0" w L1CJ fl 0v N < 0 ow Ow Z Qi— <0- 0< 00